Based on GLOBOCAN 2012 (Ferlay et al. 2015), head and neck cancer accounted for 4.8% of the 14.1 million new cancer cases worldwide in 2012, a slight decrease from 5.0% in 2008, attributable to smoking cessation programs. However, recent epidemiological data showed that there is a dramatic increase in the incidence of oropharyngeal cancer (OPC) in many western countries, in contrast to a decreasing trend in other HNSCC (Gillison et al. 2014; Chaturvedi 2012). The rapid increase of OPC is attributable to the emergence of HPV+ OPC. The global incidence of OPC was 85,000 new cases in 2012, of which 26% were HPV+ (~22,100 new cases per year) globally, and the prevalence of HPV+ OPC is much higher (50%) in developed countries (e.g., North America, Japan, and Australia) (Gillison et al. 2014; Giuliano et al. 2015). In fact, HPV+ OPC now comprises the majority of HNSCC referrals in North America. If current trends continue, it is projected that the incidence of HPV+ OPC will overtake that of HPV-related cervical cancer by the year 2020 (Chaturvedi et al. 2011).

Although WHO initially included both the oral cavity and oropharynx as potential sites of HPV-driven HNSCC, convincing evidence exists that HPV-related (HPV+) HNSCCs mainly occur in the oropharynx (predominantly in the tonsil and tongue base), while the prevalence of HPV-driven oral cavity SCC or other non-oropharynx sites is much lower than previously reported (Castellsague et al. 2016; Zafereo et al. 2016). A systematic review of published literature regarding the presence of HPV by PCR in non-oropharyngeal sites (oral cavity, larynx, nasal cavity, nasopharynx, and sinonasal SCC) reported a rate exceeding 20%. However, it is now evident that PCR-based HPV detection has overestimated oncologic-relevant HPV infection. In fact, the presence of high-risk HPV in the tumor does not necessarily indicate that HPV is a “driver” of the tumor but could be a “bystander” indicating an HPV coinfection at time of tumor detection. A study with large data on HPV DNA detection by PCR and p16 expression in HNSCC biopsies suggests that the probability of a cancer of the oral cavity, larynx, and hypopharynx being attributable to HPV is at least fivefold lower than that for oropharyngeal cancer (Combes and Franceschi 2014). A recent comprehensive analysis (Castellsague et al. 2016) of HPV biomarkers in 3680 HNSCC patients, using more robust HPV testing methods to differentiate HPV-driven tumor from coexisting HPV infection in tumor, has demonstrated that the prevalence of HPV+ tumor is less than 5% in the oral cavity, larynx, hypopharynx, and nasopharynx. A potential etiological role for HPV in a small proportion of sinonasal tumors is also suggested but confirmation is needed (Bishop et al. 2013; Syrjanen and Syrjanen 2013).

In addition, some benign lesions of the head and neck are also associated with HPV infection. For example, recurrent respiratory papillomatosis (RRP) has been confirmed as an HPV-related disease in both children and adults linked to low-risk subtypes 6 and 11 (Larson and Derkay 2010; Can et al. 2015). Malignant transformation may occur in a very small proportion of RRP with HPV subtype 11. Another potential HPV-related benign head and neck lesion is sinonasal inverted papilloma (SNIP), a locally aggressive neoplasm arising in sinonasal mucosa (Lisan et al. 2016). A growing number of molecular epidemiological studies have suggested an association between HPV infection and SNIP. Both low-risk and high-risk HPV subtypes have been detected in SNIP, but whether the presence of HPV is a “driver” or a “passenger” is yet to be fully elucidated (Lisan et al. 2016; Strojan et al. 2012; Thavaraj 2016). Studies have linked low-risk and/or high-risk HPV infection with SNIP (Buchwald et al. 1995; Hasegawa et al. 2012; Lin et al. 2016; Zhao et al. 2016; Mendenhall et al. 2007) and reported that about 5% of SNIPs could convert to invasive SCC, especially those infected with high-risk HPV subtypes.

In summary, high-risk HPV is a confirmed carcinogenic agent in oropharyngeal cancer (OPC) and may also be responsible for a small proportion of SCC beyond the oropharynx and specifically in the oral cavity, larynx, hypopharynx, nasopharynx, and sinonasal mucosa. Low-risk HPV infection is associated with head and neck benign lesions including laryngeal papillomatosis and papilloma in the sinonasal mucosa. A small proportion of the latter might experience malignant transformation, especially those infected by HPV 11 or several high-risk HPV subtypes.

HPV is a non-enveloped double-stranded DNA virus from the papillomavirus family which affects human. HPV subtypes are commonly grouped as either cutaneous or mucosal, depending on the type of tissue of prediliction. More than 170 different HPV subtypes have been identified so far, of which more than 40 are mucosal subtypes (de Villiers 2013). Based on its oncogenic (cancer-causing) ability, it can be divided into low-risk (subtypes 6, 11, 42, 43, 44, 45) and high-risk (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68) and undetermined subtypes (Kumar et al. 2015). HPV genotyping analysis has shown that more than 90% of HPV+ OPC is caused by the HPV 16 subtype and the remainder are caused by other HPV subtypes (e.g., HPV 18, 33, and 35) (Gillison et al. 2014; Castellsague et al. 2016; Kreimer et al. 2005) (see Chap. 1). A greater diversity of high-risk HPV subtypes, such as 31, 33, 35, and 49, exist in non-oropharyngeal mucosa. The genotyping variation seems to be associated with prognostic significance in disease outcomes. Recent studies suggest that HPV+ HNSCC caused by the HPV 16 subtype seems to be associated with better outcomes compared to those caused by non-HPV 16 subtypes (Bratman et al. 2016; Goodman et al. 2015). This may partially explain why HPV+/p16+ non-OPC fare less well compared to HPV+/p16+ OPC (Chung et al. 2014).

Laryngeal papillomatosis is mostly caused by HPV 6 and 11 and rarely by HPV 16 or 18 (Larson and Derkay 2010). Malignant transformation is rare. A recent analyses of 35 adult RRP patients showed the HPV was present in all patients (80% were positive for HPV 6, 8% for HPV 11, and 1% for HPV 16). Another study on RRP showed that 95% of cases were HPV positive, of which 69% were caused by HPV 6, 27% by HPV 11, and 8% by HPV 16. HPV 11-related RRP seems to behave more aggressively compared to those caused by HPV 6. Both low-risk (6 and 11) and high-risk HPV (16, 18, 33, 57, etc.) were detected in SNIP (Zhao et al. 2016), and there was a strong association of HPV 16 and 18 with malignant SNIPs (Zhao et al. 2016).

Since HPV+ HNSCC occurs predominantly in the oropharynx and the natural history and clinical behavior are much less understood in non-oropharyngeal cancer, the discussion that follows will be confined to screening, diagnosis, staging, and treatment issues in the OPC setting.

Unlike cervical cancer, HPV+ OPC often lacks a visible precancerous lesion which poses a challenge in screening (Kreimer 2014). The French “Split” trial showed that tonsil “brushing” appeared to be less reliable in detecting HPV (Franceschi et al. 2015) compared to an oral rinse method (Gillison et al. 2012), but both are not OPC specific. Recently it was shown that serum HPV16 E6 antibody was detectable in >90% HPV+ OPC patients 2–10 years prior to their cancer diagnosis (Kreimer et al. 2013b). HPV16 antibodies with at least one early protein (E1, E2, E4, E5, E6, or E7) were detected in the sera of 90.6% of HPV+ OPC cases, 0% of partners, and 7.4% of healthy volunteers (Anderson et al. 2015). The detectable E6 antibody seropositive rate is very high (approaching 100%) in HPV+ OPC patients and much lower in non-OPC (oral cavity, larynx) and genital HPV+ cancers (e.g., cervix, vagina, vulva, and penis), except in anal cancer (Kreimer et al. 2013b); it is also rarely detected in healthy individuals (Lang Kuhs et al. 2015). For this reason, it is promising to use this biomarker to design a screening algorithm for HPV+ OPC and anal cancer (Brotherton et al. 2016). However, the choice of study end point and monitoring method in seropositive patients is a dilemma since there is no visible precancerous lesion to screen, even though severe dysplasia can exist. Bilateral tonsillectomy is not an ideal follow-up procedure for E6 seropositive patients as it is invasive and associated with a 50% chance of missing HPV+ OPC in the base of tongue.

Prevention of HPV+ OPC is potentially possible by modifying sexual behavior and HPV vaccination. Tonsillectomy does not prevent HPV+ base of tongue cancer although it might reduce the risk of HPV+ tonsillar cancer. A recent study has shown that previous tonsillectomy modifies the odds of both tonsil and BOT cancer, with decreased odds of tonsil cancer and increased odds of BOT cancer (Zevallos et al. 2016).

Since there is no screening tool to facilitate early diagnosis, HPV+ OPC diagnosis largely relies on traditional clinical processes: symptom/signs prompt patients to visit medical professionals, biopsies confirm the primary tumor location, and HPV testing confirms the role of HPV.

Confirming an HPV+ OPC diagnosis can be challenging due to its unique clinical presentation and/or occasionally idiosyncratic patient behavior. Delay in diagnosis could occur when a patient fails to recognize the onset of symptoms or is in denial of a potential cancer diagnosis since many of such individuals are previously healthy without typical risk factors for traditional HNSCC. It could also occur due to clinician-related factors, such as unfamiliarity with potential initial presentations of HPV+ OPC, resulting in delayed referral from the family doctor to otolaryngologists and/or technical challenges in obtaining tissue or misdiagnosis by otolaryngologist (Lee et al. 2015; Truong Lam et al. 2016; Yu et al. 2008). Many HPV+ OPC patients present as “unknown primary” with cervical lymph node metastasis without visible mucosal lesions. An asymptomatic neck mass (typically at level 2) is the most common initial clinical presentation for about two-thirds of patients with HPV+ OPC and is often the first sign prompting medical referral (Truong Lam et al. 2016; McIlwain et al. 2014). Local symptoms, such as dysphagia or odynophagia, are much less frequent and often mild, even in the presence of a large primary compared to HPV– cases (Truong Lam et al. 2016). An HPV+ nodal mass is often of cystic appearance (Goldenberg et al. 2008; Morani et al. 2013) (Fig. 11.2) and could be mistaken for a “branchial cleft cyst.” They often have a “spongy” feeling on palpation that could be mistaken for a “lymphoma.” Cervical lymph node involvement occurs earlier in the course of HPV+ OPC development, even with small (T1–T2) primaries, likely attributable to the anatomic structure of tonsillar crypt epithelium, which has a discontinuous basement membrane and contains numerous small blood vessels (Lewis and Chernock 2014). Table 11.1 summarizes typical clinical presentations of HPV+ OPC.

Small primaries arising “deep” to the basal cell layer of crypts with minimal mucosal changes coupled with a paucity of local symptoms and cystic lymph node(s) present a technical challenge in confirming the diagnosis for this patient population. This is particularly true if there is lack of awareness that the clinical presentation of this disease is different from “classic” HNSCC caused by smoking/alcohol. For example, several cases with small but visible mucosal changes in the tongue base still require repeated biopsies to yield a diagnosis which likely occurs because the initial biopsy attempt was insufficiently “deep.” Initial lack of success in establishing a diagnosis could also be explained by difficulty in acquiring representative tissue in small HPV+ primary tumors. The presence of a cystic lymph node is also challenging in yielding positive cytology when only a small focus of solid tumor exists in the wall of the cystic lymph node.

For the specialist facing an “unknown primary” with cervical lymph node involvement fitting the correct patient profile, especially when the lymph nodes are soft or “spongy” on palpation or with a cystic component on CT and or MRI, the likelihood of HPV(+) tonsil or BOT primary should be high on the potential list of diagnoses. Obtaining tissue (e.g., core lymph node biopsy) for p16 staining would be more reliable due to established limitations of p16 staining on FNA specimens (Bishop et al. 2015). P16 positivity would highly suggest tonsil or BOT origin. PET CT may also identify a potential primary and should preferably be undertaken prior to invasive procedures to avoid false positivity arising from biopsy artifact (Huang et al. 2015a). Tonsillectomy may identify a tonsil primary more successfully compared to tonsil biopsy. Emerging evidence suggests that trans-oral robotic surgery (TORS) is a useful diagnostic tool to detect an “unknown” primary arising from the tonsil or base of tongue and is a promising therapeutic option for small tonsil/BOT primary with low-volume neck disease (Kang et al. 2015; Mehta et al. 2013). Notably, multicentric synchronous primaries may occur in a small proportion of HPV+ patients (e.g., contralateral tonsil or other head and neck regions outside the oropharynx) (Huang et al. 2012; Joseph et al. 2013; Roeser et al. 2010; Rasband-Lindquist et al. 2016; McGovern et al. 2010). Adequate pretreatment workup including clinical, endoscopic, and radiological exam is important to identify potential sites of cancer (e.g., contralateral tonsil, tongue base, supraglottic larynx). It has been reported that some HPV+ synchronous head and neck lesions may not be obvious with routine CT/MRI and fiberoptic examination and may only be successful during examination under anesthesia (Huang et al. 2012).

Histological confirmation of HPV-driven OPC is important since HPV+ and HPV– OPC have different prognoses, and staging and treatment might also differ (see discussion below) in the foreseeable future. Although HPV+ tumors are SCC, they are often poorly differentiated of nonkeratinizing or basaloid morphology.

To determine whether a tumor is HPV driven, specific HPV testing is required (Boscolo-Rizzo et al. 2016). Many HPV testing methods exist including assays for HPV E6 and E7 DNA or mRNA by PCR or in situ hybridization (ISH) . Currently there is no consensus regarding optimal tumor HPV testing. Table 11.2 summarizes HPV testing methods which can be used for the detection of HPV-positive tumors, of which p16 staining is generally accepted as a reliable surrogate marker for HPV-driven OPC, if appropriately scored and interpreted. Strong and diffuse p16 immunohistochemistry staining indicates an HPV+ OPC and does not require further confirmation, while patchy or weak p16 staining requires further HPV testing (El-Naggar and Westra 2012; Thomas and Primeaux 2012; Shi et al. 2009).

Although p16 is an acceptable surrogate marker for HPV-driven OPC, it is not a reliable marker for non-OPC cases (Zafereo et al. 2016; Chung et al. 2014; Maxwell et al. 2010). Confirming the presence of oncologically relevant HPV E6/E7 DNA in tumor is required for HPV-driven tumor in non-OPC (Bishop et al. 2015; Young et al. 2015).

Although HPV+ and HPV– OPC are two different diseases, presently they are still classified by the same TNM staging system (UICC/AJCC 7th edition TNM). Convincing evidence exists that the current TNM staging system (7th edition) (Dahlstrom et al. 2013; Huang et al. 2015b; Keane et al. 2016; O’Sullivan et al. 2016) is not adequate for HPV+ OPC. An HPV+ OPC-specific new TNM stage classification has been proposed (Huang et al. 2015b) and refined and validated in a multicenter dataset (O’Sullivan et al. 2016) which combined N1-N2b into N1 category and collapsed T4a and T4b into only a T4 category. Table 11.3 presents the TNM classification for HPV+ OPC introduced by the UICC/AJCC 8th edition TNM.