1. IntroductionThe global incidence of oropharyngeal squamous cell carcinoma (OPSCC) is estimated to be approximately 93,000 new patients per year [1]. During the past decades, the incidence of OPSCC has been increasing in the Western world [2], mainly due to an increase in human papillomavirus-positive (HPV+) OPSCCs [3,4,5]. We previously reported a high and increasing incidence of OPSCC with age-adjusted incidence rates (AAIR) per 100,000 from 1.8 in 2000 to 5.1 in 2017 in eastern Denmark, primarily driven by an increase in HPV+ OPSCC with 65% of OPSCCs being HPV+, with HPV16 as the predominant genotype (86%) [4,6,7,8,9]. In addition to HPV, well-known risk factors for developing OPSCC are tobacco smoking and alcohol consumption [10].HPV+ OPSCC patients have a better prognosis and demonstrate distinct clinical, histopathological, and genetic characteristics compared to HPV-negative (HPV-) OPSCC patients [10,11,12,13,14]. The tumour suppressor p16 is often overexpressed in HPV+ OPSCCs, and may therefore be used as a surrogate marker for HPV DNA. Solely double positivity for p16 and HPV DNA (p16+/HPV DNA+) implicates a better prognosis [15].Secondary primary cancer (SPC) is the second most common cause of death among patients with head and neck squamous cell carcinomas (HNSCCs), including OPSCC [16,17,18]. SPCs are divided into two groups: synchronous SPCs diagnosed at the same time or within six months after the primary cancer and metachronous SPCs diagnosed more than six months after [19]. The majority of SPCs are metachronous, while synchronous cancer is a relatively rare event but is clinically important since it might affect treatment regimen [20].Alcohol consumption and tobacco smoking are notable risk factors for developing SPC, especially regarding cancers localized in the upper aerodigestive tract, lungs, and oesophagus [21]. Studies have shown that p16+ OPSCCs have lower risk of SPC [22], primarily driven by a decrease in SPCs within the upper aerodigestive tract [23,24].

It is important to better characterize the risk of SPC in OPSCC and the relation to HPV status to improve screening, treatment, and follow-up strategies and hence oncologic outcomes.

To our knowledge, the incidence and pattern of SPC in patients with OPSCC have not been evaluated in a population-based study in a high-HPV-prevalence area in patients with known HPV DNA and p16 status. The aim of this study was to examine the risk of SPC in patients with OPSCC in eastern Denmark from 2000–2020. Further, we evaluated the survival following SPC in relation to HPV status.

4. DiscussionTo our knowledge, this is the largest consecutive, population-based retrospective study including 2584 patients diagnosed with OPSCC that all were HPV- and p16-tested. In this cohort from eastern Denmark from years 2000–2020, we found a high risk of SPC following a primary OPSCC. The risk was nearly five times higher (IRR = 4.96) compared to the risk of any cancer (the same type of malignancies) in the general population in Denmark [28]. It is expected that head and neck cancer patients are more likely to develop SPCs since they smoke and drink more compared to the general population [30]. This applies especially to patients diagnosed with SPC in the upper aerodigestive tract, the lungs, or the oesophagus, where the exposure of these carcinogens is most concentrated. Our study supports this theory since most patients with SPC were former or current smokers, and significantly more patients with SPC had a smoking history compared to patients without SPC. León, X. et al. previously found an increased risk of SPC for patients continuing tobacco smoking and alcohol consumption after treatment of a head and neck cancer, including OPSCC [31]. Whether patients continued to smoke tobacco and consume alcohol or had quit after their primary OPSCC diagnosis was unfortunately not reported in this study, but the previous finding by León, X. et al. [31] should be used in guidance of patients to improve prognosis after an OPSCC diagnosis.We found an increased risk of SPC in older patients, which is not surprising since age is a well-established risk factor in all types of cancer. We also saw a higher risk of SPC in patients with a UICC8 stage II OPSCC at the time of diagnosis, which calls for further investigation since we cannot describe a clear explanation. Patients with a N2 classification at the time of diagnosis of OPSCC were shown to have a significantly lower risk of SPC in our analysis. This might be because HPV+ patients more often have lymph node involvement at the time of diagnosis. It would be of interest to examine baseline characteristics stratified by HPV status to further investigate the observed association. The risk of SPC was not associated with T-stage OPSCC or sublocation. HPV-positivity showed to be a significant protecting factor, which is in line with the before-mentioned differences in baseline characteristics between the subgroups of HPV+ and HPV- OPSCC patients. Further, Bosshart, S.L. et al. [32] also found HPV+ to be a significant protecting factor for SPC, which substantiates our finding. However, they did not find an association between either age or tumour stage and the risk of SPC. The last-mentioned inconsistency and the association between tumour stage and the risk of SPC following an OPSCC calls for further research. Most patients with OPSCC received primary radiation therapy with or without chemotherapy, which is a well-known risk factor for developing SPC, but according to Hashibe, M. et al. [33], the earliest radiation-induced SPCs for solid tumours are detected 10 years after treatment. Since we only had four patients diagnosed with SPC in the head and neck region later than ten years after the index tumour, we do not consider radiation as the major cause of SPC in this study.The overall predominant sites of SPC were the lungs, head, and neck, and the gastrointestinal tract. The SPC localizations differed between patients with HPV+ and HPV- OPSCC index tumours and head and neck were only the third most frequent site for HPV+ OPSCC compared to the second most frequent site in HPV- OPSCC. In contrast to patients with HPV- OPSCC, gender-specific SPCs were more common for patients with HPV+ OPSCC than head and neck malignancies. Further, HPV+ patients had a better overall survival than HPV- patients, reflecting the well-established clinical differences between patients with HPV+ and HPV- OPSCC. Since the lungs are the most predominant site for both HPV+ and HPV- and since the lungs are a common site for distant metastasis from OPSCC, we find it relevant, in the future, to further studies to investigate new malignancies in the lungs following OPSCC to better differentiate between SPC and recurrence [34]. Further, the different distribution of SPC localizations could be linked to an immunosuppressed state, which was described in patients with HPV+ OPSCC earlier [35,36]. Interestingly, the incidence of SPC within the LHN was significantly different in HPV+ and HPV- OPSCC (p = 0.003), at 37.6% and 55.0%, respectively. We also found that patients with SPC outside the LHN area had a significant and better survival compared to patients with SPC within the LHN (p = 0.01), which probably reflects the inferior survival of lung cancer and the challenges of treating a second cancer in the head and neck area. Holstead, R. et al. [23] also found an association between HPV+ OPSCC and SPC outside the LHN area and requested further investigation for development of individualized guidelines. Due to the high risk of SPC following OPSCC, clinicians should be aware of this in the post-treatment surveillance of patients, especially in the first years after diagnosis (median time to SPC 2.0 years). The attention should be further increased in HPV- patients since they develop SPC significantly earlier than HPV+ OPSCC patients (p = 0.001) and have a worse survival, which is in line with other studies [21,24]. Additional, especially the LHN area should be investigated close for early detection of SPC to improve survival.According to HPV status and the overall risk of SPC, we found a significantly lower risk of SPC in patients with HPV+ OPSCC index tumours. Other studies also found a lower risk of SPC in patients with p16+ OPSCC [22,32] but did not use double positivity for both HPV DNA and p16, as in this study. Based on the same database, Grønhøj, C. et al. [34] established earlier that HPV+ patients are less likely to have both distant- (p = 0.02) and loco-regional (p = 0.001) recurrence following OPSCC. In this study, we found a significantly better overall survival following SPC for HPV+ OPSCC compared to HPV- (p = 0.003). These findings could be due to the patients diagnosed with HPV+ OPSCC being younger, demonstrating less comorbidity, and having a more favourable smoking history, as shown earlier [37]. Furthermore, this might also reflect the differences in SPC localization in HPV+ OPSCC and HPV- OPSCC patients, with the incidence of SPC within the LHN area being significantly lower in the HPV+ OPSCC group. Most patients in this study were treated with radiotherapy or chemoradiotherapy, which challenges treatment of a SPC in the head and neck area since a second radiation in the same area is often not an option. This might explain the inferior survival in patients with SPC in the head and neck area, which is shown to be more represented in HPV- patients and might support their poorer survival.The current study was restricted by its retrospective design. Further, there is no international consensus on how to define an SPC, and hence, the definition differs between studies [3,20,21,31,38]. In this study, we considered all second tumours within five years after primary OPSCC diagnosis at the same localization as a recurrence of the index tumour though it could be a potential SPC. This might affect the total number of SPC. A worldwide restricted consensus on the definition of SPC could be useful when comparing studies.

The strength of this study was the large size of the population-based, non-selected, consecutive cohort of all patients diagnosed with OPSCC in eastern Denmark with a high number of events, appropriate follow-up time, and a broad knowledge of clinical data and tumour characteristics, including both HPV DNA and p16 status. Further, it was a strength that all patients received a standardized treatment through the whole period, following the same guidelines.