4. Epidemiology – Oropharyngeal and Hypopharyngeal Carcinoma

The field of epidemiologically defined risk factors (except for HPV16-associated OPSCC) is still very limited for oropharyngeal and hypopharyngeal carcinoma. In addition to the above-mentioned factors of tobacco and alcohol, there are other factors, but these take a back seat [25][26][27] Similarly, there are neither specific mutations nor molecular-histological subtypes, such as in breast cancer, that allow a prognostically differentiated classification. The risk profile for oropharyngeal and hypopharyngeal carcinomas identified in epidemiological studies is predominantly very similar to the risk factors for oral cavity carcinoma [1]. Chronic tobacco or alcohol abuse increases the risk of disease up to 6-fold, and a combination of both risk factors up to 30-fold. In addition to the consumption of tobacco or alcohol, an unbalanced diet, such as excessive consumption of meat or fried food, can also increase the risk of carcinoma developing in the oral cavity and throat region [28][29]. Conversely, it has been shown that a balanced Mediterranean diet more than halves the risk of developing carcinoma in the throat, adjusted for nicotine consumption and BMI. The key protective elements of the Mediterranean diet are citrus fruits, vegetables (especially fresh tomatoes), olive oil and fish oils [30][31].

A few reports make a connection with individual sectors or occupational groups. Tobacco and alcohol-adjusted case-control studies and cohort studies have consistently described an association between employment in the construction industry, among painters and varnishers and in metalworking occupations and the occurrence of throat cancer. The relative risks or standardized mortality rates range between 1.5 and 3. In individual studies, a correlation was also found for employees in the paper and rubber industry. The studies on the textile industry and the woodworking trades showed inconsistent results [32][33][34][35]. It is therefore necessary to take a detailed occupational, nicotine and alcohol history in patients with throat carcinomas in order to determine the significance of occupational and non-occupational causation. Careful consideration of the various risk factors in individual cases will make it possible to identify patients with pharyngeal carcinomas in whom occupational exposure is likely to be an equally important partial cause of the disease. This requires cooperation between the attending physician and an occupational physician. A definitive compensable occupational disease for throat cancer (compared to larynx and paranasal sinuses) has not yet been defined in the Occupational Diseases Ordinance (BKV).

Malignant tumours of the oropharynx and hypopharynx are 95% squamous cell carcinomas. Most squamous cell carcinomas have different degrees of differentiation. Grading is carried out according to the worst differentiated part of the tumour. However, the grading anchored in the WHO only has very limited prognostic value [36]. The extent of keratinization is also considered to have little prognostic value [37]. Newer concepts of tumour grading of laryngeal carcinomas show a higher predictive value (see below).

The updated WHO classification has been available since 2017, in which a number of innovations and concepts have been included with regard to squamous cell carcinomas and their precursors. The etiologically defined classification is the main change in the new concept. HPV-associated and toxin-triggered squamous cell carcinomas are clearly differentiated as independent tumour entities, whereby HPV-associated oropharyngeal carcinomas are no longer graded according to the conventional scheme as before and have been given their own ICD number in the WHO classification (ICD-O 8085/3). A two-stage system (low-grade vs. high-grade dysplasia) has been proposed for the precursors of squamous cell carcinoma [38] [39].

Subtypes of squamous cell carcinoma

Squamous cell carcinomas (PECA) are divided into subtypes according to the WHO classification. In addition to PECAs with classic morphology, there are special forms such as verrucous PECA (also known as Ackermann's tumour). This is highly differentiated (G1) and exhibits a so-called pushing border phenomenon ("displacing growth"). In addition, basaloid squamous cell carcinomas are explicitly differentiated (by definition: G3, high-grade). These have a worse prognosis than conventional squamous cell carcinomas and are usually advanced at the time of diagnosis. The histological hallmark is basaloid differentiation with tumour cells arranged in a pallisade-like pattern at the edge of the tumour nests. These tumours should not be confused with HPV-associated PECAs, which can also be associated with a basaloid and non-keratinizing morphology. Papillary and verrucous squamous cell carcinomas are prognostically more favourable due to their superficial growth, but are very rare tumours overall (1 - 4%). Verrucous squamous cell carcinoma can have focal infiltration foci and then behaves like a conventional squamous cell carcinoma. Spindle cell carcinoma (formerly: sarcomatoid carcinoma, carcinosarcoma) can arise de novo or after radiation from a conventional squamous cell carcinoma [36].

Grading is based on nuclear pleomorphism and architecture. HPV-associated carcinomas are not (or are no longer) graded, because the conventional morphology is often G3. In the future, further classifiers (formation of tumour buds, so-called budding or formation of tumour cell separation) could be used to assess the grading. Good prognostic accuracy has been reported [36] [38] [39].

Precancerous precursor lesions

Non-invasive precursor lesions of squamous cell carcinoma are referred to as epithelial dysplasia and, according to the 2005 WHO classification, synonymously as intraepithelial neoplasia (squamous intraepithelial neoplasia: SIN). They are classified as low-, moderate- and high-grade (SIN 1-3); an important criterion here is, among other things, the disruption of the epithelial architecture in the lower, middle or upper third. In the new nomenclature of intraepithelial neoplasia, no distinction is made between carcinoma in situ and high-grade intraepithelial neoplasia: High-grade intraepithelial neoplasia (SIN 3) = carcinoma in situ. The malignancy risk of low- and moderate-grade intraepithelial neoplasia (SIN 1 and SIN 2) is 11%, while that of high-grade intraepithelial neoplasia (SIN 3 = carcinoma in situ) is 90%. It is not necessary for a lesion to pass through all stages of intraepithelial neoplasia to become a squamous cell carcinoma. Squamous cell carcinoma can arise from all grades of intraepithelial neoplasia and even in morphologically inconspicuous mucosa [37].

The nomenclature of intraepithelial lesions is subject to constant change (see other organs such as the cervix or intestine), with three or two-stage systems essentially being used: SIN I-III vs. low and high grade dysplasia [36]. It may be advisable to specify both terms (neoplasia and dysplasia) in the text of the findings.

In both previously published S3 guidelines on laryngeal carcinoma and oral cavity carcinoma, overview texts were prepared on histological precursor lesions as risk factors, which apply equally to oropharyngeal and hypopharyngeal carcinoma [1][2].

For oropharyngeal carcinoma (OPSCC), it is noticeable that the "classic" risk factors of tobacco/alcohol consumption have been overshadowed by the now prominent and reliably substantiated causal role of infection with human papillomavirus (predominantly high-risk subtype HPV16) (particularly for tonsil and tongue base carcinomas, which are the most rapidly increasing head and neck subsites). It is now assumed that HPV-associated OPSCC is a genetically diverse tumour subgroup distinct from HPV-negative oropharyngeal carcinomas [40][41][42][43]. In particular, the early expressed ("early", E) proteins E5, E6 and E7, which are encoded by the viral genome, contribute to this. The oncoprotein E7 binds and destabilizes the retinoblastoma tumour suppressor protein (pRb), releasing factors that are necessary for transcription, proliferation and cell cycle progression. As a by-product of this interaction, the protein ^p16INK4A (hereafter p16) is highly expressed [44][45]. In infected cells, E6 leads, among other things, to inactivation of p53 and thus to the prevention of cell cycle control, thereby increasing genetic instability. Tumours that are purely HPV-associated often show a histological phenotype reminiscent of a basaloid squamous cell carcinoma, but without belonging to the subgroup of basaloid squamous cell carcinomas in the narrower sense. HPV proteins also lead to "immune escape", which makes chronic infection and thus possible malignant degeneration more likely [46][46][47], although a single infection with HPV does not necessarily lead to malignant degeneration. E5 supports the expression of growth factors and epidermal growth factor receptor (EGFR) and thus increases cell proliferation. However, EGFR expression is usually reduced in HPV-positive tumours [48]. In HPV-driven tumours, the p53 wild type is usually found and no TP53 mutations, which are associated with tumour development by classical noxious agents.

Since the corresponding morphology is subject to a certain range of variation, histological classification alone is unreliable. The detection of HPV16 mRNA E6*I, a sequence coding for the neoplastic transformation-causing proteins E6 and E7, is currently regarded as the most reliable detection method for definitive HPV16 association; however, it is often difficult to implement in routine diagnostics [24][49].

Importance of p16 as an HPV16 surrogate parameter

The most common method currently used to "detect" an HPV infection is p16 immunohistology (see Chapter 7 for specific diagnostics). If squamous cell carcinomas strongly express p16, this is indicative of an HPV association of OPSCC. However, up to 23% of p16-positive OPSCCs that are examined using polymerase chain reaction (PCR) and in situ hybridization are ultimately HPV-negative [Mehanna, H et al. 2023] [Rischin, D et al. 2010][52]. This applies in particular to sites (e.g. larynx) where HPV cancers are rather rare and for suboptimal material [53] as well as in the context of a non-HPV16 HPV association. The frequency of p16-positive and HPV-DNA-negative tumours is lower in regions with a high incidence of HPV-associated tumours. Despite this uncertainty, p16 is currently the simplest and cheapest method for indirect HPV16 detection and is therefore unanimously recommended by the AJCC and UICC TNM Committee. In routine clinical practice, p16 immunohistology followed by HPV DNA detection (PCR or in situ hybridization) has proven effective in identifying oropharyngeal carcinomas that are truly HPV-associated.

HPV16 transmission, "high risk sexual behaviour", geographical differences

HPV transmission occurs predominantly through skin and mucous membrane contact. It is currently assumed that HPV is primarily transmitted through sexual intercourse (genital, anal, oral), but for oncogenic and non-oncogenic HPV types, contact with public wet surfaces (toilets, door handles, public pools, etc.) is also likely. Sexual behaviour in particular is viewed differently as an infection risk factor in various epidemiological studies. In 2007, Maura Gillison's group published (New England Journal of Medicine) the until recently undisputed observation that the HPV-16-associated risk of developing oropharyngeal carcinoma is particularly associated with high-risk sexual behaviour. From a total of >26 "self-reported" vaginal sex partners (high-risk sexual behaviour; HR-SB) over the entire lifespan, a highly significant association with the occurrence of oropharyngeal cancer was observed (odds ratio 3.1; CI 1.5-6.5), which correlated with the increasing number of partners. A comparable odds ratio was also calculated for the number of >6 "self-reported" oral sex partners [D'Souza, G et al. 2007]. Brenner et al. [55] essentially describe risky sexual behaviour (i.e. a higher number of sexual partners, same-sex sex, younger age at sexual debut) as risk factors for the occurrence of early HPV antibodies. In addition to the factors mentioned above, men between the ages of 51 and 60 were also described as a high-risk group in the USA [56]. In a recent publication from the Leipzig LIFE cohort (propensity score matching 112 oropharyngeal carcinoma patients with 303 controls from the normal population), the association with HR-SB could not be confirmed, at least in Germany (greater Leipzig area). No differences were seen in the self-reported number of lifetime vaginal and oral sex partners between the Leipzig propensity score-matched sample of oropharyngeal cancer patients and controls. A comparison of the Leipzig results with the above-mentioned study by D'Souza shows a significantly lower prevalence of HR-SB in control subjects and an even lower prevalence of HR-SB in oropharyngeal carcinoma cases. The consistent absence of HR-SB in the overwhelming majority of HPV-related oropharyngeal cancers is also presented in light of a lower frequency of HPV-related oropharyngeal cancers in the Leipzig cohort compared to observations in the US (35.1% versus 64%, corresponding to seropositivity for HPV16 E6 and/or E7 antibodies).

In fact, there is increasing evidence that does not allow a generalized transfer of epidemiological data from the USA to Europe, especially Germany, without a more differentiated view. We see these differences in the evaluation of HR-SB as a risk factor and in the consideration of the discordance of the value of p16 as a surrogate parameter for a genuine HPV16 involvement in the development of oropharyngeal carcinoma (for more details, see chapter 6.2 [Mehanna, H et al. 2023]). The HPV type spectrum is also different: the spectrum in Europe is narrower and focussed on HPV16 compared to a broader spectrum in the USA, which also includes other types, such as HPV18/45, which we generally do not see in Europe.

The HPV-16 virus plays an almost exclusive role in the genesis of oropharyngeal carcinomas. Infection with the HPV viruses occurs in the basal cells of the squamous epithelium, and the lymphatic tissue in the tonsils (crypt epithelium) enables the viruses to gain access to the basal cells even without injury. In the other parts of the mucosa, infection is only possible in the case of erosion or minor trauma. Infections can heal or lead to a latent infection. [24][49] [57][58].

Critically, numerous authors in Germany point out that purely HPV16-associated oropharyngeal carcinoma, which is not triggered by noxious agents, is very rare in Germany. The vast majority of patients (not exactly quantifiable) have a mixture of p16 positivity and existing noxious agent exposure. In this respect, the boundaries between the two oropharyngeal carcinoma entities, which are supposedly different, are blurred.

It is possible that the smoking habits of patients are a major factor responsible for the geographical impact on HPV prevalence rates since then. It appears to be well established that patients with HPV-positive OPSCC are predominantly non-smokers, particularly in the USA and Canada. Therefore, in countries with a comparatively small proportion of smokers, such as those regularly described for US cohorts, HPV16 prevalence is significantly higher than in regions with a higher proportion of smokers, such as Germany. For example, HPV16 prevalence in Sweden (the lowest proportion of smokers (7%) in Europe) is estimated at 70%. In Germany, where the proportion of smokers is 24%, HPV16 prevalence is estimated at only 40%. The described interaction between smoking habit and HPV status is not fully understood. However, smoking appears to have a protective effect against the cancer-causing HPV16 infection. Based on the results of the Kiel group around Hoffmann M and Quabius ES on more than 1000 patients and supported by two US-American studies, the following hypothesis is currently being discussed: Smoking leads to significantly increased "secretory leukocyte protease inhibitor" (SLPI, an antileukoproteinase) and AnxA2 (annexin A2) expression in mucosal tissue. SLPI, which is excessively elevated in smokers, binds to AnxA2, which in this combination prevents the binding of HPV, if present. The binding of HPV to AnxA2 is crucial for a successful HPV infection of the mucosal cells. Conversely, in non-smokers with significantly higher AnxA2 levels, HPV can bind more easily to unoccupied - non-SLPI-bound - AnxA2, making successful infection of the cells more likely [59][59]. The promoting influence of cannabis (marijuana) on the development of HPV16-positive OPSCC has also been proven (through p38 MAPK pathway activation) [60][Marks, MA et al. 2014].