6. Prognosis, predictors – Oropharyngeal and Hypopharyngeal Carcinoma

For oropharyngeal carcinoma (OPSCC), it is noticeable that the "classic" risk factors of tobacco/alcohol consumption have been overshadowed by the now prominent and sufficiently substantiated causal role of infection with the human papillomavirus (predominantly high-risk subtype HPV-16), 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 subgroup distinct from HPV-negative oropharyngeal carcinomas with a significantly better prognosis [40],[41] [42][86][43].

Retrospective analyses suggest that HPV-associated OPSCC responds significantly better to previous treatment concepts. This tumour group showed significantly better survival after primary surgical as well as radio- or radiochemotherapy or anti-EGFR treatment. HPV is not a predictor for a specific treatment regimen and the prognosis is comparable after primary surgery plus adjuvant treatment and primary radiochemotherapy in retrospective cohorts [87]. In the majority of studies, overall survival after 5 years averaged up to 80% for HPV-positive (or p16-positive) OPSCC and 30-35% for HPV-negative OPSCC [24].

The retrospective analyses of an American study published in 2010 (RTOG 0129), in which accelerated fractionated radiotherapy in combination with cisplatin was compared with standard fractionated radiotherapy with cisplatin, provided the key impetus for HPV16-dependent therapeutic considerations. Kian Ang (a highly respected radiotherapist from the M.D. Anderson Cancer Center USA, who was instrumental in establishing the clinical consideration of HPV16 and unfortunately died unexpectedly in 2013) [8] was able to develop the following trend-setting 3-tier score(the good, the bad, the ugly) depending on tobacco consumption, HPV-16 status, N-status and tumour size:

1. Low-risk tumours: HPV-16-positive, ≤ 10 pack years or >10 pack years + N0-N2a
2. Intermediate-risk tumours: HPV-16-positive + > 10 Pack Years + N2b-N3 or HPV-16-negative + ≤ 10 Pack Years + T2-T3
3. High-risk tumours: all other HPV-negative

(UICC 7th Edition)

The HPV-16-positive never smokers differed from the HPV-16-negative heavy smokers in terms of 5-year survival by almost 50% [8] (note: the 7th edition of the TNM classification was used here). However, the stratification proposed by Ang does not differentiate in a comparable way in all retrospective cohorts. In the case of HPV-positive tumours, low tobacco consumption was not relevant to prognosis in various studies.[88]

In contrast, oropharyngeal (HPV-16/p16 negative) and hypopharyngeal carcinomas are squamous cell carcinomas of the upper aerodigestive tract, which have almost identical risk factors and often premalignant precursor lesions (for an overview, see Section 4.3.3). In analogy to laryngeal carcinoma, it can be postulated for oropharyngeal and hypopharyngeal carcinoma that the prognosis essentially depends on the localization, TNM classification and R status. Furthermore, differentiation and the presence of lymph vessel invasion (lymphangiosis carcinomatosa) are relevant to prognosis. Oropharyngeal carcinomas have better prognoses than hypopharyngeal carcinomas.

In both previously published S3 guidelines on laryngeal carcinoma and oral cavity carcinoma, overview texts were prepared on premalignant lesions, risk factors and prognostic factors, which apply equally to oropharyngeal and hypopharyngeal carcinoma (for an overview, see the S3 Guideline for Oral Cavity Carcinoma 007-100OL, 2021, Ch.7.4; for an overview, see S3 Guideline for Laryngeal Carcinoma 017-076OL; Ch.3.2.1; 4).

The p16 protein is encoded by the CDKN2A gene, is a CDK (cyclin-dependent kinase) inhibitor and plays a central role in cell cycle regulation and cellular senescence [89], [90]. In noxious-associated head and neck tumours, p16 is often inhibited by methylation, mutations or deletions (Network CGA). In contrast, in HPV-associated OPSCC, tumour cells are usually characterized by strong p16 overexpression, which is caused by the activity of viral HPV oncoproteins, especially E7, and can be used as a surrogate parameter for this entity [44][45][91] [43][92].

Therefore, p16 detection via immunohistochemical staining (p16-IHC) with FFPE tumour sections is recommended as a suitable surrogate parameter for HPV status in OPSCC in routine diagnostics according to the current edition of the TNM classification (AJCC 8th Edition) and the CAP (College of American Pathologists) and ASCO (American Society of Clinical Oncology) guidelines [93][94][92]. For positive detection, at least 70% of the tumour cells should show heterogeneous, nuclear and cytoplasmic staining with moderate to strong intensity. Comparative studies of IHC staining with commercial anti-p16 antibodies and FFPE tumour sections showed moderate differences in specificity, intensity and variability as assessed by independent observers [95] [96].

Patients with a p16+/HPV16- OPSCC have a significantly worse prognosis compared to p16+/HPV+ tumours [Mehanna, H et al. 2023][97] [98] [99] [100][101] and show a slightly better risk profile than HPV-/p16- tumours [Mehanna, H et al. 2023][102]. In a recent pooled analysis of 13 studies across Europe (7895 patients), Mehanna et al showed that the 5-year overall survival for patients with p16+/HPV16+ status was 81.1% (95% CI 79.5-82.7), for p16-/HPV16- 40.4% (38.6-42.4), for p16-/ HPV16+ 53.2% (46.6-60.8) and for p16+/HPV16- 54.7% (49.2-60.9). Patients with discordant OPSCC (p16-/HPV+ or p16+/HPV-) had a significantly worse prognosis than patients with p16+/HPV+ OPSCC and a significantly better prognosis than patients with p16-/HPV16- OPSCC. In the collectives contributed from Germany (1035 patients from Cologne, Giessen and Kiel), the proportion of p16-/HPV16+ was 1.4-13.1%, p16+/HPV- 2.7-9.4% [Mehanna, H et al. 2023]. Thus, additional HPV-specific detection via viral DNA or RNA in combination with p16-IHC appears to be urgently required for the inclusion of patients in clinical trials [93] [94]. When reporting the test result, it should be stated whether the detection was carried out using p16-IHC as a surrogate parameter (p16-positive) or an HPV-specific method (HPV-positive) [93]

Since the corresponding morphology is subject to a certain range of variation, histological classification alone is unreliable for reliable HPV detection in the narrower sense. 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 not possible in routine practice (fresh material) [24]. In fact, relevant working groups report 15-29.7% of p16-positive OPSCCs that were ultimately HPV16-negative in the polymerase chain reaction (PCR) and in situ hybridization [Mehanna, H et al. 2023][Rischin, D et al. 2010][52]. The discordance p16+/HPV16- was highest in oropharyngeal sublocations outside the tonsils and tongue base [Mehanna, H et al. 2023].

The detection of HPV-16 mRNA E6*I is therefore currently regarded as the most reliable detection method for the definitive HPV-16 association in oropharyngeal carcinomas.

The sensitivity and specificity of p16-IHC as a surrogate parameter for the presence of an actual HPV association has been investigated in numerous studies. A meta-analysis of 24 studies showed a sensitivity of 94% (95%-CI: 91-97%) and a specificity of 83% (95%-CI: 78-88%) for p16-IHC alone compared to the detection of viral E6/E7 transcripts as a reference [103]. A high sensitivity but moderate specificity in the diagnosis of HPV-positive OPSCC has been confirmed in more recent studies, where the proportion of p16+/HPV16- OPSCC is between 4-29% [93] [94][97] [98] [99].

Holzinger et al [104] were able to show that sensitivity and specificity can be significantly increased by combining p16 and pRb ICH, with increased p16 expression in combination with underexpression of pRb (78% sensitivity, 93% specificity, 78% PPV, 93% NPV). The vast majority of retrospective studies therefore emphasize p16 and establish the association with HPV-16 as equivalent despite the above-described uncertainty. Not least for this reason, therapeutic conclusions based on retrospective analyses should be weighed very carefully. In the 8th edition of the TNM classification, however, only p16 is (currently still) required as a discriminator.

In addition to p16-IHC as a surrogate parameter, there are several established methods for the detection of HPV-positive OPSCC with tumour tissue. These are based on the detection of viral DNA or RNA via in situ hybridization (ISH) or PCR-based technologies using fresh tumour samples or tissue sections [99]. The detection of transcripts of the viral oncoproteins E6 and/or E7 is often used as a reference or gold standard for the detection of HPV-positive OPSCC with tumour tissue [105]. However, the detection of viral transcripts is technically more demanding and more expensive, which makes it difficult to implement in current routine diagnostics [105] [94]. Modern ddPCR (droplet digital PCR) and NGS (next generation sequencing) based technologies enable the non-invasive detection of circulating HPV DNA in blood or saliva [99]. Several prospective studies have shown a high concordance between HPV detection in blood and tumour tissue and demonstrate the potential of circulating HPV DNA in assessing treatment success and monitoring the risk of recurrence in primary HPV-positive OPSCC as well as recurrences and metastases [106][107][108] [109][Mattox, AK et al. 2022] [111]. In a direct comparison with blood and saliva samples from 66 patients with HPV16-positive OPSCC, NGS technology showed the highest sensitivity in the detection of HPV DNA compared to ddPCR or quantitative RT-PCR. However, these data as well as the practical implementation in clinical routine need to be confirmed and verified in larger, standardized clinical studies [99] [107][111].

The revision of the TNM classification for oropharyngeal carcinomas has been fundamentally revised, particularly on the basis of the ICON-S study [O'Sullivan, B et al. 2016] [Huang, SH et al. 2015] [Haughey, BH et al. 2016] [93]. The 8th edition of the TNM classification (UICC, AJCC) is currently available, in which a different, p16-associated approach with a reclassification of tumour stages and N-status has been introduced. The anatomical assignment and definition of the regions on which the TNM classification is based was explained in detail in Chapter 3.

Therefore, if an OPSCC is p16-positive, not only pT1N0M0 tumours are subsumed under stage I, but now also pT1.2 pN0.1 tumours. Stage II even allows pT1.2 pN2 and pT3 pN0.1 tumours, compared to only pT2N0 previously; stage III includes pT3 pN2 and pT4 pN0-2 (8th edition TNM 2017). The staging of p16-positive OPSCC is differentiated clinically and pathologically. cN2 is stage II and pN2 is stage III. It should also be noted that the N category is categorized according to the number of positive lymph nodes (pN1: 1-4; pN2: ≥ 5) and thus differs from all other squamous cell carcinomas and a tumour-perforated lymph node capsule (extracapsular extension, ECS) [Huang, SH et al. 2020] is no longer taken into account. Several authors complain about the resulting purely classificatory "down-staging", which was justified and enforced in the 7th edition of the TNM classification on the basis of the poorer stage discrimination of p16-positive oropharyngeal carcinomas that had previously only been observed retrospectively. Furthermore, clinically and pathologically different TN stages were introduced. The changes were therefore made with a view to purely prognostic significance. When considering tumour stages, great care must therefore be taken not to fall into the danger of "under treatment", especially in the case of parallel tobacco, alcohol and HPV associations [116] [117][118]. The TNM classification of hypopharyngeal carcinoma has not changed since the 7th edition.

The histopathological report has already been described in detail in the S3 guideline on laryngeal carcinoma and is only supplemented by the p16 findings in this guideline [2].

The pTNM classification as the basis for the pathological assessment of resectates has already been explained above. The treatment decision is also influenced by the following parameters, some of which are also associated with or depicted in pTNM:

• Tumour location and size,
• Histological tumour type according to the current WHO classification [39], [119] [38],
• Local tumour extent, infiltrated structures,
• Lymph node metastases separated according to level (siege figure neck levels according to Robbins) and side:
• Number of lymph nodes examined,
• Number of affected lymph nodes,
• Largest diameter of the lymph node metastases,
• Tumour growth beyond the capsule,
• Lymph vessel/venous invasion and perineural invasion,
• Presence of an in situ component (with size),
• Differentiation of the tumour according to the established grading scheme [120],
• Distance to the lateral and basal resectate margins for all relevant resection margins and for the invasive and in situ components in mm.
• R classification
• Indication of the p16 expression status (positive, negative). For positive detection, at least 70% of the tumour cells should show heterogeneous, nuclear and cytoplasmic staining with moderate to strong intensity [93][94][92].

For biopsies, samples should be taken from the periphery of the tumour and, if possible, centrally from the base of the tumour. Resectates for the pathologist should be provided with topographically clear information regarding the anatomical orientation and exact localization.

The corresponding parameters mentioned above must be precisely stated in the histopathological description of the findings (where applicable). For biopsies, the number of parameters is naturally reduced.

Biopsies:

Clinically, biopsies of a manifest macroscopically clear neoplasm should be taken from the periphery of the tumour and, if possible, centrally from the base of the tumour (see also below). The examination order to the pathologist must contain all clinically relevant information. If the findings are unclear, the biopsy should be repeated after consultation with the pathologist.

Resectates:

In the case of resectates, the tumour specimen should be sent to the pathologist by the surgeon with a clear description of the anatomical topography (suture or color marking, needle marking, clock dial diagram) (BDP2017). In case of doubt, a personal consultation should take place. In the case of a neck dissection, the levels must be marked separately or sent in portions as individual specimens for pathohistological examination. The cut edges on the resectate can be examined by means of a frozen section according to clinical relevance, so that resection can be carried out in the same session in the event of tumour involvement. It is preferable to send the resectate "en bloc" and have the surgeon or pathologist take a frozen section of the resectate margins. Areas with specific clinical questions should be marked separately. This procedure preserves the integrity of the specimen in the best possible way and allows the most reliable statement on the R status, as small sections do not have to be virtually adapted to each other retrospectively.

Transoral robotic surgery (TORS) or laser microsurgery (TLM) resections of primary tumours of the oropharynx and hypopharynx are increasingly used approaches for cancer resection in selected patients with accessible tumours. Oncologic principles are similar to open procedures. Therefore, if safe en bloc resection is not possible ("piecemeal" technique in the context of transoral surgery, TLM or TORS), it is appropriate to send separate margin sections, which should ideally also be sent to the pathologists as frozen sections. In this case, the tumour fragments should be marked on cork and arranged in relation to each other according to the resection. The transtumoural section margins should be marked separately from the outer areas relevant as margin sections. The subsequent frozen sections should be clearly assigned to the resection margins. For this purpose, it is advisable to invite the pathologist to the operating theatre and agree on a documentation concept together. This makes it possible to determine a final R status even after piecemeal surgery. The previous refusal in principle of some pathologists to give an R status after piecemeal surgery, who then classified RX, should be overcome in the context of good interdisciplinarity.

Cervical lymph nodes:

Locoregional metastasis of the primary tumour to the cervical lymph nodes is a reliable negative predictor of prognosis, with the more lymph nodes involved, the less favourable the course of the disease. Furthermore, involvement of the caudal levels (IV and V) and growth beyond the capsule have a negative impact on the prognosis [121][121][122][123][124] [125] [126] [127] [128]. It is preferable for the surgeon to portion the levels, as exact ex-situ allocation in pathology is only possible to a very limited extent, even when markings are applied. The dissection of lymph nodes (especially those with tumour involvement) should be avoided. The lymph nodes should be assigned to the level at which the largest lymph node diameter is visualized. Structures of particular interest should be marked separately.

The histopathological findings of the neck dissection specimens should include the side of the neck, the cleared levels, the total number of lymph nodes with the number of affected lymph nodes per level, the diameter of the largest lymph node metastasis, additionally removed structures and - if present - information on growth that exceeds the lymph node capsule (extra nodal extension, ENE). As the current discussion on the different extensions of ENE has not yet resulted in any clinical implications, this aspect is currently not addressed further in this guideline. The detection of isolated tumour cells in lymph nodes or of micrometastases, which can only be detected using immunohistology, is currently still of unclear clinical relevance [129].

For further explanations regarding the histopathological assessment of oropharyngeal and hypopharyngeal carcinomas, please refer to the current S1 guideline on the pathological anatomical diagnosis of squamous cell carcinomas of the head and neck, 1st edition (BDP2017).

[1][2].

An overriding goal of oncological surgery is complete tumour resection with histological evidence of tumour-free margins. The margins can be assessed in real time intraoperatively by frozen section or by assessing formalin-fixed tissue. Tumour-free margins are a basic principle of surgical strategy to reduce the risk of local tumour recurrence. Conversely, positive margins increase the risk of local recurrence and are an indication for postoperative adjuvant therapy [130]. Classic clinical-pathological studies with unchanged current relevance have shown the significance of narrow or positive margins and their correlation with local tumour recurrence [131]. Other authors only observe a negative influence on freedom from local recurrence with R1, but not with any safety margins in mm from the tumour border in the case of an R0 resection [130]. The removal of frozen sections is generally recommended. In a recent multicentre Italian  study (450 patients, 10 centres), it was even shown that in patients in whom R0 resection was verified intraoperatively by frozen sections, the local control rate was twice as high as in those without frozen section verification. This is a surprising, unexpected and not easily explained result. The authors suggest that frozen section assessment is a quality indicator of surgical and pathological procedures and their effectiveness [130].If there is an initial incision through an invasive tumour at the surgical margin, additional harvested adjacent margins from the patient (i.e. not from the excised tumour) may mean a higher risk of local recurrence and should therefore be described in the surgical report. When obtaining additional margins from the patient, there may be confusion as to whether the tissue removed from the surgical bed corresponds to the actual location of the positive margins [132]. If positive surgical margins are reported, a repeat resection should be considered if possible or, if not reasonably possible, adjuvant therapy should be considered. The patient's prognosis is not worsened by an initial R1, or close R0 situation, which leads to a final stable R0 situation through resection at the correct site [133][134] [130].

Assessment of frozen section margins is always at the surgeon's discretion and should be considered if it facilitates complete removal of the tumour. Achieving sufficiently wide margins may require resection of an adjacent structure in the oral cavity or laryngopharynx, such as the base and/or anterior tongue, mandible, larynx or part of the cervical oesophagus.

Adequate resection is defined by current international consensus as clean resection margins with at least sufficient distance from the gross tumour to obtain a clear frozen section and permanently free margins (can often mean 1.0-1.5 cm of visible and palpable normal mucosa). In general, a frozen section examination of the margins is usually performed intraoperatively and when a resection line is not clear due to indistinct tumour margins or when residual disease is suspected (i.e., soft tissue, cartilage, carotid artery or mucosal irregularities). For transoral endoscopic and robotic approaches for oropharyngeal and hypopharyngeal cancers, margins of 1.5-2.0 mm may be acceptable, but data are based on retrospective studies, so caution is advised [135]. Such margins would in principle be considered "narrow or closed" and are considered too narrow for certain sites, such as the base of the tongue.

• A clear resection margin is defined as the distance from the invasive tumour front that is 5 mm or more from the resected margin. There is international agreement on this, which is reflected in the current NCCN guidelines [136]. For hypopharyngeal carcinomas with diffuse, partially submucosal spread, particularly towards the oesophagus, the 5 mm safety margin should be regarded as the absolute minimum and can be generously extended to 1 cm for safe resection according to international consensus.
• A narrow margin is defined as the distance from the invasive tumour front to the resected margin that is less than 2-5 mm depending on the anatomical site affected.
• A positive margin is defined as carcinoma in situ or invasive carcinoma at the resected margin. If carcinoma in situ is present and if additional clean margins can be achieved by resection, this is the preferred approach. Carcinoma in situ should not be considered an indication for concomitant postoperative systemic therapy/RT.

If the surgeon takes additional resection margins from the patient, the new margins should relate to the geometric orientation of the resected tumour specimen with a statement from the pathologist that this will be the final resection margin and its histologic status in the final tender (NCCN Guideline Version 2.2022, SURG-A 4 of 8).

The unfavourable prognostic impact of pretreatment (especially surgical pretreatment) for local control regardless of R0 resection in the context of salvage surgery has been pointed out [130]. Tumour grading (often criticized as inconsistent and subjective) also appears to have a greater prognostic role than previously assumed. In the aforementioned multicentre Italian study, the group of well-differentiated squamous cell carcinomas showed a statistically significant and clinically relevant difference in 5-year tumour relapse-free survival and OS compared to the group of poorly differentiated squamous cell carcinomas, depending on localization and independent of R0 resection. [130]