Introduction

Squamous cell carcinoma (SCC) of the UADT associated with the presence of the HPV displays distinct clinical features that affect the presenting signs and symptoms of disease and can be used to help clinicians in diagnosis and work-up. These distinctive clinical features constitute powerful evidence for an active role of HPV in the etiology of HPV-related SCC.

Site specificity

One of the most specific and distinct features of HPV-related SCC is the region of the UADT in which it arises. HPV-related HNSCC arises predominantly in the lymphoepithelium of the OP in tissues commonly called tonsils. The tissue known as tonsils by the lay public is the palatine tonsils. The target of HPV-related tumorigenesis, however, includes both the palatine tonsils and the lingual tonsils. In a seminal article, Gillison and colleagues showed that of 253 HNSCCs, all those with integrated HPV capable of tumor induction, were located in the OP, and oropharyngeal positive tumors arose predominantly from the palatine or lingual tonsils. Surface epithelial areas within the OP are not commonly associated with HPV-related SCC. Tumors of the uvula, soft palate, tonsillar pillars, and posterior pharyngeal wall do not typically contain HPV DNA. It may be difficult, however, to accurately identify the site of origin of large tumors that involve multiple subsites.

Controversy exists in the literature as to whether HPV-HNSCC arises in other sites within the UADT. Reports of a small number of laryngeal and oral cavity (OC) SCC lesions that contain HPV DNA appear occasionally. There may be several explanations for this:

• 1.
  

  HPV may truly affect mucosa outside the oropharyngeal lymphoepithelium under some undefined rare circumstances.

  
  
• 2.
  

  Laboratory methodology, such as polymerase chain reaction (PCR), used to identify HPV in some reports may have lower thresholds that result in categorizing a tumor as HPV positive when the amount of HPV signal is below biologic significance, indicative of laboratory background levels or simple infection without cancer transformation.

  
  
• 3.
  

  The true site of origin may be inaccurately recorded in the medical record. In the author’s experience, this third explanation is often uncovered on closer examination of the medical record of cases with HPV-related tumor assigned to a non-OP site of origin. For example, the author found several cases in which samples taken from the oral tongue were found to be HPV-related HNSCC; however, the past history indicated that a tonsillectomy had been performed at an outside facility before the tongue biopsy was obtained and the tonsil specimen contained SCC. These cases are rightly categorized as having an OP primary site. The actual location of tissue taken from the vallecula or inferior lateral oropharyngeal wall may be called supraglottic larynx or hypopharynx, respectively, in cases where the epicenter of tumor is actually the tongue base or inferior palatine tonsil.

The reliance on surgical pathology reports alone to categorize the site of origin of a cancer case is problematic because specimens may be labeled in a manner that does not fully describe the orientation and extent of the entire tumor. Inexperience or lack of precision may lead a surgical endoscopist to mislabel the site from which a specimen is taken. Similarly, reports from computerized imaging may be misleading in that a radiologist is not always able to accurately locate a tumor boundary due to the contiguity and overlapping of structures in the UADT. The preponderance of a particular site of origin of HPV-related HNSCC in the OP in the majority of cases from virtually every report in the literature supports the author’s contention that many if not all outlier cases could be explained by 1 of the 3 alternatives discussed previously.

One possible exception to this site specificity, which seems to be emerging, is the occasional presence of HPV-related SCC arising in the nasopharynx. Investigators at the University of Michigan reported that 4 cases of nasopharyngeal carcinoma in white subjects enrolled in a clinical trial had tumors that tested positive for HPV. Recently, Singhi and colleagues reported several cases in which nasopharyngeal biopsies were found to contain HPV whereas Epstein-Barr virus was more commonly associated with nasopharyngeal carcinoma. In all cases with available staging information, a tonsil lesion was also present.

The reason for the site specificity of HPV-related cancer to the tonsillar tissue of the OP is not fully understood. Lymphoepithelial tissue architecture seems designed to permit contact between external pathogens and the immune system. A permissively porous basement membrane has been described. In addition, the deep crypts and irregularity of both the palatine and lingual tonsils may permit prolonged exposure of the epithelium with entrapped food debris and pathogens, which are protected from the purging of mucosal lining that regularly occurs elsewhere in the UADT with swallowing. There may be additional factors that link these physical features to explain the link between exposure, infection, and tumorigenesis.

Site specificity

One of the most specific and distinct features of HPV-related SCC is the region of the UADT in which it arises. HPV-related HNSCC arises predominantly in the lymphoepithelium of the OP in tissues commonly called tonsils. The tissue known as tonsils by the lay public is the palatine tonsils. The target of HPV-related tumorigenesis, however, includes both the palatine tonsils and the lingual tonsils. In a seminal article, Gillison and colleagues showed that of 253 HNSCCs, all those with integrated HPV capable of tumor induction, were located in the OP, and oropharyngeal positive tumors arose predominantly from the palatine or lingual tonsils. Surface epithelial areas within the OP are not commonly associated with HPV-related SCC. Tumors of the uvula, soft palate, tonsillar pillars, and posterior pharyngeal wall do not typically contain HPV DNA. It may be difficult, however, to accurately identify the site of origin of large tumors that involve multiple subsites.

Controversy exists in the literature as to whether HPV-HNSCC arises in other sites within the UADT. Reports of a small number of laryngeal and oral cavity (OC) SCC lesions that contain HPV DNA appear occasionally. There may be several explanations for this:

• 1.
  

  HPV may truly affect mucosa outside the oropharyngeal lymphoepithelium under some undefined rare circumstances.

  
  
• 2.
  

  Laboratory methodology, such as polymerase chain reaction (PCR), used to identify HPV in some reports may have lower thresholds that result in categorizing a tumor as HPV positive when the amount of HPV signal is below biologic significance, indicative of laboratory background levels or simple infection without cancer transformation.

  
  
• 3.
  

  The true site of origin may be inaccurately recorded in the medical record. In the author’s experience, this third explanation is often uncovered on closer examination of the medical record of cases with HPV-related tumor assigned to a non-OP site of origin. For example, the author found several cases in which samples taken from the oral tongue were found to be HPV-related HNSCC; however, the past history indicated that a tonsillectomy had been performed at an outside facility before the tongue biopsy was obtained and the tonsil specimen contained SCC. These cases are rightly categorized as having an OP primary site. The actual location of tissue taken from the vallecula or inferior lateral oropharyngeal wall may be called supraglottic larynx or hypopharynx, respectively, in cases where the epicenter of tumor is actually the tongue base or inferior palatine tonsil.

The reliance on surgical pathology reports alone to categorize the site of origin of a cancer case is problematic because specimens may be labeled in a manner that does not fully describe the orientation and extent of the entire tumor. Inexperience or lack of precision may lead a surgical endoscopist to mislabel the site from which a specimen is taken. Similarly, reports from computerized imaging may be misleading in that a radiologist is not always able to accurately locate a tumor boundary due to the contiguity and overlapping of structures in the UADT. The preponderance of a particular site of origin of HPV-related HNSCC in the OP in the majority of cases from virtually every report in the literature supports the author’s contention that many if not all outlier cases could be explained by 1 of the 3 alternatives discussed previously.

One possible exception to this site specificity, which seems to be emerging, is the occasional presence of HPV-related SCC arising in the nasopharynx. Investigators at the University of Michigan reported that 4 cases of nasopharyngeal carcinoma in white subjects enrolled in a clinical trial had tumors that tested positive for HPV. Recently, Singhi and colleagues reported several cases in which nasopharyngeal biopsies were found to contain HPV whereas Epstein-Barr virus was more commonly associated with nasopharyngeal carcinoma. In all cases with available staging information, a tonsil lesion was also present.

The reason for the site specificity of HPV-related cancer to the tonsillar tissue of the OP is not fully understood. Lymphoepithelial tissue architecture seems designed to permit contact between external pathogens and the immune system. A permissively porous basement membrane has been described. In addition, the deep crypts and irregularity of both the palatine and lingual tonsils may permit prolonged exposure of the epithelium with entrapped food debris and pathogens, which are protected from the purging of mucosal lining that regularly occurs elsewhere in the UADT with swallowing. There may be additional factors that link these physical features to explain the link between exposure, infection, and tumorigenesis.

Occult nature of early lesions

One clinically relevant effect of the origination of HPV-associated HNSCC in palatine and lingual tonsils is the occult nature of early lesions. Deep within tonsillar crypts or hidden amidst lingual tonsil fronds and mounds, early mucosal alterations due to HPV malignant transformation may grow undetected for many months. This is different than the early manifestation of cancers of the floor of mouth and ventral tongue in the OC. It is these OC tumors that are commonly associated with progression of leukoplakia and erythroplakia progressing from atypia to dysplasia. These OC lesions are also associated with field cancerization, with multifocal tumors arising particularly in heavy smokers/drinkers. Field cancer effects are not seen in HPV-associated HNSCC, and precancer lesions in the tonsil and tongue base are rarely, if ever, reported. This is also different from the HPV-related precancer lesions seen on the uterine cervix and detected by Papanicolaou smear. The routine screening and accessibility of the cervical mucosa for testing of atypical cells and HPV DNA may facilitate early detection of precancerous changes—similar tests are not yet routinely available for OP screening.

OP tumors may grow to considerable size before causing symptoms, such as pain, dysphagia, or referred otalgia. Tonsillar asymmetry may be the presenting complaint, leading to a diagnosis of HPV-associated HNSCC. Finding a tumor in the tongue base is more difficult. The tongue base falls vertically away from the view of an examiner of the oral cavity, making it impossible to visualize without special equipment even for an interested examiner and a patient with a controlled gag reflex. When a laryngeal mirror or fiberoptic laryngoscope is used, the tongue base is still viewed tangentially, making thorough inspection difficult. Using a fiberoptic laryngoscope, examination is facilitated by maneuvers, such as asking the patient to protrude the tongue or puff the cheeks. The difficulty in visualizing and examining the tongue base and tonsil crypts contributes to a prolonged interval between presentation of an individual to a health care provider with complaints related to oropharyngeal cancer and the eventual diagnosis of the tumor. Although the interval between presentation and diagnosis is typically up to 9 months, there does not seem to be a negative impact on outcome of treatment attributable to this interval.

Palpation of the tongue base is a vital part of the physical examination for early detection of base of tongue (BOT) tumors. Many care providers avoid BOT palpation fearing that a patient may vomit, bite, or refuse examination. The author finds that with proper preparation (explanation of the importance, encouragement to breathe steadily, and warning that patients will experience gag), BOT palpation is accepted and effective. This has been a standard part of the physical examination of the head and neck for many decades. What is being investigated is whether there are any discrete areas of increased firmness or raised mass effect in the tongue base. Examination in the clinic is limited not only by the gag reflex but also by tongue muscle turgor. A better evaluation is possible when patients are under general anesthesia with muscle relaxant in effect. Under these conditions, the extent of a tumor can be more easily assessed compared with the soft consistency of the tongue base musculature. Circumvallate papillae can feel firm and should not be mistaken for tumor. The size of these raised plaque-like taste organs is characteristically small (3 mm to 5 mm) and can be visualized while palpated to confirm their identity.

The site specificity of HPV-associated HNSCC is particularly relevant when searching for the primary tumor in cases where lymph node metastasis provides the initial histologic evidence of the cancer. If a node biopsy has confirmed HPV-associated cancer, the site of the primary tumor is very likely within the palatine or lingual tonsil. This is discussed in depth when UP cancer is considered.

Nodal disease

The presenting sign of HPV-associated HNSCC frequently is the abrupt appearance of an enlarged cervical lymph node, which has implications in the diagnosis and evaluation of the disease. There are several other clinical details about HPV-associated HNSCC metastases that should be thoroughly understood. There are factors that contribute to the common phenomenon of a lymph node as the first and often only clinically apparent manifestation of HPV-associated HNSCC. First, the individuals affected often do not have the classical lifestyle risk exposures associated with HNSCC, specifically, extensive smoking and alcohol use, so are not considered at risk for the disease. Often people are not even aware that HNSCC exists before their own personal encounter. Another factor is the occult, symptom-free clinical course of many OP primary tumors. Many individuals who present with a metastatic lymph node enlargement report no problems related to the mouth or throat. It seems that HPV-associated oropharyngeal tumors often remain small in the face of large cervical metastases. Again, several factors may be at play: there may be a differential rate of tumor cell growth between primary and metastatic disease if the latter arises from a more transformed and more rapidly growing subclone of the tumor; metastatic tumor may have greater access to nutrients through proximity to blood vessels; and perhaps the predominant reason for the rapid appearance of large neck nodes in HPV-related HNSCC is the phenomenon of the cystic metastatic node.

For many years, the phenomenon of a cystic mass of recent onset presenting in the neck of young to middle-aged adults in whom SCC is found has been described. Reports in the literature of SCC arising in a branchial cleft cyst have been published and debated. The mechanism for malignant transformation of a benign epithelial cyst has been the subject of speculation. More recently, the observation that HPV-related SCC is often associated with cystic lymph node metastasis has been made. A cystic lymph node is described as having a thin wall, containing clear liquid ( Fig. 1 ). Because these nodes are often reported of recent abrupt onset, the mechanism for their growth may be obstruction and accumulation of lymphatic flow rather than growth of neoplastic cells beyond available blood supply with subsequent necrosis and liquification. In that case, the body of the mass is more irregular with areas of solid tumor and collections of caseous necrotic material.

Cystic cervical lymph node metastasis in HPV-related HNSCC.

The association of cystic nodal metastasis related to HPV-HNSCC is now so commonly recognized that the historical phenomenon of SCC arising in a branchial cleft cyst may be best reinterpreted as an early example of HPV-related disease. More importantly, any adult presenting with a new onset of a lateral cystic cervical mass should be suspected of having HPV-related HNSCC until proved otherwise. Lateral cystic neck masses appearing abruptly in adults should be worked up by fine-needle aspiration (FNA) with PCR of the aspirate looking for HPV DNA. If benign squamous epithelium is present, the most likely diagnosis remains metastatic SCC and further work-up is warranted, including a direct laryngoscopy with directed biopsies. Ultrasound guidance may help identify areas of solid nodal metastases that coexist with a cystic node. If surgical extirpation is undertaken, the mass should be cleanly dissected, taking care not to violate the capsule, and the mass submitted for frozen histopathologic analysis. In anticipation of a cancer diagnosis, adequate operating room time should be requested and consent obtained for both completion neck dissection (at least regional, selective neck dissection) and examination under anesthesia with directed biopsies of the oropharyngeal lymphoepithelium in search of the primary tumor site. Failure to anticipate the likely diagnosis of cancer and to take these measures frequently results in suboptimal oncologic surgery with possible spillage of tumor cells and a need for revision surgery within a scarred field, putting vital structures, such as cranial nerves, at risk as well as losing an opportunity to complete the surgical diagnostic work-up under a single anesthetic.

A review of cystic cervical metastatic nodes at Johns Hopkins Hospital revealed that the majority of cases were from an HPV-related tonsil/tongue base primary. Papillary thyroid carcinoma may also occasionally spawn cystic nodal metastasis. HPV-associated cystic nodes commonly occur, however, in the level II-III junction of the upper lateral neck to mid-lateral neck and achieve a larger size than papillary thyroid carcinoma cystic nodal metastases. Not all HPV-related SCC cervical nodal metastases are cystic; most are solid and commonly cystic and solid nodes coexist in the same patient.

Cystic metastasis may help explain two clinical phenomena: the abrupt appearance of large neck masses as a presenting sign of HPV-related HNSCC and an improved response rate to treatment of nodal metastasis in HPV-related HNSCCs compared with the formerly more common HNSCCs arising from exposure to tobacco and alcohol. In the first observation, rapid accumulation of lymphatic fluid may occur when a tumor implant reaches a size to block the effluent channels within a node. Fluid accumulation within a node may occur more quickly than would tumor parenchymal growth, consistent with the report of many patients that a node suddenly appeared and was noted, for example, while shaving. Similarly, if the substance of a nodal metastasis is largely lymphatic fluid, the size on which nodal staging is based is inflated and disproportionate to the actual tumor burden. There are orders of magnitude fewer viable tumor cells in an 8 cm ³ (2 × 2 × 2 cm) cystic node compared with a solid nodal metastasis of the same size. The responsiveness of HPV-related HNSCC to various treatments seems better than that of non-HPV–related cancer, ostensibly because of a less disrupted genetic milieu in tumors that undergo malignant transformation because of the effect of the E6 and E7 HPV oncoproteins. Stage inflation, however, in which nodal metastases with small tumor volumes are graded as N2 or N3 disease, may also contribute to an apparent improvement in clinical outcome. This phenomenon may have a role in the current debate about the necessity for a “planned” neck dissection in which a neck dissection is done after treatment regardless of clinical response when the presenting disease stage is N2 or above. This dogma may be less warrented for HPV-related HNSCC.

The position of cervical nodal metastasis from HPV-related HNSCC primary disease is also highly characteristic. As discussed previously, the deep jugular nodal chain in the upper-mid neck is the typical location for these metastases. Nodal disease that appears more anteriorly, even along the common facial vein between levels I and II, is suggestive of a non-HPV–related source. Both the palatine and lingual tonsils can spawn contralateral nodal metastases. In most cases, this is in conjunction with ipsilateral metastases but can occur as an isolated event, which is important in informing the strategy of looking for an occult primary tumor as well as in design of the scope of neck treatment (radiation or surgery) required in these cases. Another consideration is the propensity for nodal metastasis from the palatine tonsil to spread to retropharyngeal nodes. These lymph nodes are difficult to approach surgically and are not typically removed at the time of a comprehensive nodal dissection. Again, the potential for nodal metastasis to an area behind the posterior pharyngeal wall may be indication to include the tonsil region in radiation portals even when a tonsil primary tumor has been completely resected.