In this large population-based proportional study using the SEER database, we found that HPV-positive rates were higher in male, Caucasian patients with tonsillar OSCC and age between 45 and 49 years. The number of HPV (+) and (−) patients with OPSCC did not differ significantly in 2010. However, increasing proportional trends of patients with OPSCC were only observed in not HPV (−) group but HPV (+) group since then. Analyzing past trends of HPV (+) and (−) patients with OPSCC, we could predict the trends of patients with OPSCC in the future.

In a previous large cohort study, tonsillar OPSCC showed a more favorable prognosis than non-tonsillar OPSCC in HPV (+) patients [10]. One systemic review found that lymphoepithelial sites of the oropharynx, such as the tonsil and tongue base, showed significantly higher HPV prevalence than non-lymphoepithelial sites of the oropharynx [11]. Our study showed that HPV positivity rates were significantly higher in the tonsillar area than in the oropharyngeal area in OPSCC. HPV-positive rates were the highest in the tonsillar and NOS areas in patients with OPSCC. We believe that further subdivision and investigation of the primary tumor location would help better predict the prognosis of patients with OPSCC.

HPV is a well-known cause of cervical, oropharyngeal, vulvar, vaginal, penile, and anal cancer [12]. The availability of prophylactic HPV vaccines has provided powerful tools for primary prevention of cervical cancers since the HPV vaccine was approved by the US Food and Drug Administration in 2006 for women and 2009 for men [13, 14]. Cervical cancers are preventable with HPV vaccine uptake and systemized screening, but other HPV-associated cancers including OPSCC do not have any screening guidelines. From 2001 to 2017, women showed the decreasing incidence of cervical cancer and stable incidence of OPSCC. However, men showed significant increase in incidence of OPSCC [14]. This may be because of significant differences by sex and anatomic sites as well as HPV vaccine policy [15]. Men seem to remain at high risk of HPV infection regardless of age. Additionally, the second peak in high-risk oropharyngeal HPV infection prevalence may occur at any age at men. It is necessary to observe the proportional trends of OPSCC for decades, because the HPV vaccine for men has been introduce since 2009, and the period is still too short for its effect to appear.

Data from the SEER database between 1973 and 2004 showed that the proportion of potentially HPV-related oral squamous cell carcinoma (OSCC) increased steadily, particularly among white men and at younger ages [16]. Potentially HPV-unrelated OSCC declined in general between 1973 and 2004. However, these previous data do not contain information about HPV status; thus, tumors arising from the oropharynx, including the tonsil and tongue base, were defined as HPV-related OSCC. A recent large population-based epidemiologic study using the SEER database between 2013 and 2014 showed that most HPV (+) patients with OPSCC were white males younger than 65 years [17]. In the present study, we specifically selected patients with OPSCC who had tested for HPV status in the SEER database between 2010 and 2017. We also found that the HPV-positive rates were higher in middle-aged white males. However, patients older than 70 years or of African American descent showed high HPV-positive rates (> 50%). A possible explanation for this finding could be that most OPSCCs in this study were tonsillar and that the HPV-positive rates in patients steadily increased every year. A large hospital-based national cohort study confirmed that the prevalence of HPV is increasing among male and female patients from White, Black, and Hispanic ethnicities [18].

The ARIMA model has been widely used in time-series analysis of various cancers [19, 20]. We also used the ARIMA model to predict the trends of HPV (+) and (−) OPSCC. To obtain appropriate ARIMA models for HPV (+) and (−) OPSCC, we first decomposed a time series into trend, seasonal, cyclical, and irregular components [9]. In our study, a seasonal tendency was not observed in both HPV (+) and (−) OPSCC; thus, the multiplicative seasonal ARIMA model was not applied [7]. The optimal ARIMA (p, d, q) model was confirmed using the auto.arima function of R 4.2.0 software. Through this function, we found optimal ARIMA (2,1,0) and (0,1,1) models for HPV (+) and HPV (−) patients with OPSCC, respectively. To check the accuracy of each model in the prediction of HPV (+) and (−) patients with OPSCC, 95% CIs were described. Some outbreaks were observed during the prediction of HPV (+) and (−) patients with OPSCC in 2015, but the forecast of HPV (+) and (−) patients matched the actual data in 2016.

Our study has the following limitations. First, we conducted a retrospective analysis of data from the SEER database, noting the absence of smoking and alcohol history, as well as molecular characterization. A systematic review and meta-analysis indicated a negative interaction between smoking and alcohol in the development of HPV (+) OPSCC [21]. In a previous study involving 157 OPSCC patients, gene alterations associated with oxidative stress were observed more frequently in HPV-positive cases, whereas those related to the p53 signaling and cell cycle control pathways were more prevalent in HPV-negative cases [22]. Furthermore, it revealed significant disparities in immune characteristics, including B cells, CD8 + T cells, fibroblasts, and M2 macrophages, among HPV-positive cancers across different anatomical sites, such as the oropharynx, non-oropharynx, and cervix [23]. Further investigation into these demographic and molecular factors is imperative for a deeper understanding of proportional trends in HPV (+) OPSCC.

Second, this study only included patients with OPSCC diagnosed between 2010 and 2017. As this study explored HPV positivity in OPSCC cases, which are already known to be related with HPV infections, we only utilized data from 2010 to 2017 because most SEER databases from the 1990s and the 2000s contain no information about HPV status. The proportion of patients who performed the HPV testing steadily increased every year. From 2010 to 2014, the HPV testing rates increased significantly (21.95% to 51.37%), thus it is thought that the number of both HPV (+) and (−) patients with OPSCC increased. From 2014 to 2017, the HPV testing rates increase to a relatively small extent (51.37% to 58.14%), and at this time, the number of HPV (−) patients with OPSCC remained unchanged. However, the number of HPV (+) patients with OPSCC continued to increase regardless of similar testing rates between this period. These phenomena suggest proportional trends in HPV (+) and (−) patients with OPSCC.

Additionally, the coronavirus disease 2019 (COVID-19) pandemic, first diagnosed in December 2019, has posed a significant threat to global public health [24]. A single-institution retrospective study indicated a delay in the diagnosis of HPV (+) OPSCC due to the COVID-19 pandemic, factors including restricted access to medical centers and unavailability of clinic appointments [25]. However, the SEER databases do not include information on the HPV status of OPSCC patients since 2017. Consequently, the impact of the COVID-19 pandemic could not be analyzed using SEER databases. Further studies utilizing a large population database that includes HPV status information up to recent periods are necessary to comprehensively analyze the impact of the COVID-19 pandemic on the proportional trends in HPV (+) OPSCC.

Third, information on HPV status is site-specifically disproportionate as HPV tests are often conducted only in areas where HPV is suspected to be positive. This is particularly relevant for tonsillar OPSCCs. In this study, the HPV testing was performed in 50.02% and 42.96% of patients with OPSCC located in the tonsil and BOT, respectively, and the positivity rates of HPV were 78.09% and 73.97% in patients. On the other hand, the HPV testing was performed in 23.17% of patients located in SP, and the positivity rates of HPV was 21.78%. Thus, the HPV testing is required to be carried out in consideration of the tumor subsites and other proportional factors of patients with OPSCC.

Nevertheless, to the best of our knowledge, this is the first study to predict future HPV trends by observing past proportional trends of OPSCC patients using a large population database. Examining the changes in HPV status in OPSCC raises the importance of preventive approaches. Appropriate screening strategies based on various factors are important to prevent the occurrence of OPSCC. Early cancer detection with HPV testing, commercialization of vaccination, and public awareness of causative factors other than HPV status are key factors in preventing OPSCC. Further studies, including more detailed epidemiological data, will help identify other factors affecting HPV status in OPSCC.