Assessing the burden of HPV-related head and neck cancers in mainland China: protocol of a nationwide, multisite, cross-sectional study Authors: Wei Wang, Cheng Song, Zheng Su, Smita Kothari, Ya-Ting Chen, Yin Liu, Shu-Yu Wu, Ravi Panchal, Edith Morais, Shao-Kai Zhang, Jian Yin, You-Lin Qiao, Craig Roberts Published: 2023 ABSTRACT BackgroundPersistent human papillomavirus (HPV) infection is a known cause of a subset of head and neck cancers (HNCs). In the last two decades, the proportion of HNCs attributable to HPV infection has increased worldwide, notably the oropharyngeal cancers. However, the trend of HPV-related HNC burden is not clearly understood yet in China. Thus, the absolute burden of HPV-related head and neck cancers in China (BROADEN-China) will be conducted to estimate the proportion of HNCs attributable to HPV infection, per anatomic site, by genotype, in three time periods (2008–2009, 2013–2014 and 2018–2019).Methods and analysisBROADEN-China is a nationwide, multisite, cross-sectional study. A stratified, multistage, non-randomised cluster sampling method will be used to select 2601 patients with HNC from 14 hospitals across seven regions, based on population density in China. Patients with formalin-fixed paraffin-embedded tissue samples collected prior to treatment induction during three time periods will be included, and factors (eg, smoking status, alcohol consumption, betel nut chewing, Epstein-Barr virus, teeth loss, etc) associated with HNC will be assessed. HPV testing (HPV-DNA, HPV-mRNA and p16INK4a immunohistochemistry) and histological diagnosis of the tissue samples will be conducted at a central laboratory.The study protocol and all required documents have been submitted for review and approval to the Independent Ethics Committees of all the participating sites. The informed consent was waived for all participants and all the recorded data will be treated as confidential.We have included 14 hospitals as our participating sites, of which Henan Cancer Hospital is the leading site. The study has been approved by the independent ethics committees of the leading site on 3 December 2020. The other 13 participating site names of ethics committee and IRB that have approved this study. STRENGTHS AND LIMITATIONS OF THIS STUDY Burden of HPV-related head and neck cancers in China (BROADEN-China) is the first nationwide, multisite, cross-sectional study to assess the human papillomavirus (HPV)-related head and neck cancer (HNC) burden. Burden of HPV-related head and neck cancers in China (BROADEN-China) is the first nationwide, multisite, cross-sectional study to assess the human papillomavirus (HPV)-related head and neck cancer (HNC) burden. A stratified, multistage, non-randomised cluster sampling method used to select participants across seven regions based on population density will ensure the national representativeness. A stratified, multistage, non-randomised cluster sampling method used to select participants across seven regions based on population density will ensure the national representativeness. This study will report the HPV-attributable fractions across three different time periods. This study will report the HPV-attributable fractions across three different time periods. The use of archived samples to perform central HPV testing limits the eligible HNC population to those patients for whom a pretreatment tissue sample was collected and is available. The use of archived samples to perform central HPV testing limits the eligible HNC population to those patients for whom a pretreatment tissue sample was collected and is available. Introduction In the Global Cancer Statistics 2020 report, it was estimated that there were 878 348 new head and neck cancer (HNC) cases and 444 347 deaths worldwide.1 In China, there were an estimated 128 000 new HNC cases, with over 65 000 related deaths,2 with tobacco and alcohol use among the leading risk factors for HNC. Human papillomavirus (HPV) is a common sexually transmitted virus which can infect epithelial cells, alter cell cycles and ultimately cause cells to undergo oncogenic transformation. In recent years, an increase in the incidence of HPV-related HNC cases has been noted in several high-income countries, such as the UK, the USA, Italy, Norway and Canada.3–6 It was reported that the majority of HPV-related HNC was due to HPV type 16.7 The genetic, epidemiological and clinical characteristics of HPV-related HNC have been shown to be significantly different from the non-HPV-related HNC,5 6 8 which further led to varied survival outcomes. Currently, the HPV-related HNC could be potentially prevented through HPV vaccines and screening. In the USA, the non-valent HPV vaccine has been approved to prevent oropharyngeal cancer (OPC) in 2020.9 Until today, very few studies have been published in China on the HPV-related HNC disease burden, with the evidence to date limited by small sample sizes and with geographical heterogeneity.10–16 In one pooled analysis of data from years 2006–2018, HPV16 prevalence was reported as approximately 24.7% in HNC,17 with rates of 14.5%–37.0% in different regions of China. More specifically for each anatomical site, other studies showed that HPV positivity is approximately 31.6% in oropharynx, 16.2% in oral cavity,18 7.6% in larynx,18 19 6.5% in nasopharynx and 5% in hypopharynx.11 Data from years 2008–2012 revealed that the incidence of OPC was higher in the urban areas compared with rural areas, and higher in the Eastern region compared with the Western and Central regions of China.20 However, there is a lack of national population-based analysis in mainland China to explore the latest HNC incidence time trends, and HPV-attributable fraction (AF) by HNC anatomical site, genotype and regions. In 2021, a global observational study, the absolute burden of HPV-related head and neck cancers (BROADEN) study, has been initiated in five European countries (Germany, France, Italy, Spain and Portugal) and two Asian countries (Japan and China), which will examine the HPV-AFs in HNC and relevant risk factors. As China has unique epidemiological characteristics of HNC and clinical practice, therefore, a tailored design with implementation adaptations requires to be applied in mainland China. In this study, we aim to describe the unique designs and implementation practice in BROADEN-China, as complementary to the BROADEN-Global study design. BROADEN-China will provide robust estimates of HPV prevalence in HNC over the last 10-year period, which provides evidence on the HPV-related disease burden in both females and males and in disease areas beyond cervical abnormalities. Methods and analysis Study design and objectives BROADEN-China is an observational, multicentre, nationwide, cross-sectional study across three time periods (2008–2009, 2013–2014 and 2018–2019) as shown in figure 1. BROADEN-China was initiated in September 2020 and will end in December 2023. The main objective is to estimate the fraction of HNC in China that is attributable to HPV infection per anatomic site and HPV genotypes. This study will also describe patient characteristics (sex, age, location of head and neck primary tumour, histological diagnosis, stage at diagnosis, smoking habits and alcohol consumption at diagnosis, and comorbidities at diagnosis, that is, Epstein-Barr virus (EBV) infection for nasopharynx) among patients with HNC, according to HPV status. The study will further explore the time trend of HPV-AFs among the three different time periods (2008–2009, 2013–2014 and 2018–2019). Study design of BROADEN-China. BROADEN-China, burden of HPV-related head and neck cancers in China; EBER-ISH, Epstein-Barr encoding region in situ hybridisation; HNC, head and neck cancer; HPV, human papillomavirus; IHC, immunohistochemistry. The inclusion and exclusion criteria, HNC identification algorithm and HPV testing are aligned with the BROADEN-Global publication.21 Briefly, we will include over 18-year-old patients with primary HNC with pretreatment specimens available for HPV testing during the defined study period. We used a list of International Classification of Diseases, 10th revision (ICD-10) codes to define the HNC diagnosis for this study (online supplemental table 1). We will exclude those with only fine needle aspiration samples from metastatic lymph nodes, low-quality samples, or the absence of invasive cancer, and those included in the exclusion ICD-10 codes (online supplemental table 2). Patient identification will be performed in a retrospective consecutive way, starting from the patients with the most recent samples in each time period. Laboratory testing consists of SPF10-DNA enzyme immunoassay (DEIA)-line probe assay (LiPA25) (SPF10-DEIA-LiPA 25) genotyping, HPV mRNA testing (E6*I) and p16INK4a immunohistochemistry (IHC) to determine the attribution of HPV in HNC. Patients will be considered having HPV-related HNC if at least two tests among HPV-DNA, p16INK4a or HPV-mRNA are positive. Notably, Epstein-Barr encoding region in situ hybridisation (EBER-ISH) will also be performed to determine the absence of EBV related to nasopharyngeal cancer (NPC). In total, 2601 patients are estimated to be enrolled over the three time periods (2008–2009, 2013–2014 and 2018–2019). The sample size was calculated based on AF reported for other Asian countries, or the global AF where no Asian data were available. Table 1 describes sample size calculation by primary tumour site. Based on these estimations of sample size for the HPV-AF, the target number of patients to be included is 821 patients with HNC per time period. It is estimated that 5% of tissue samples will be not valid for evaluation and as per the exclusion criterion, therefore the sample size per period increases from 821 to 867. According to the population of each region reported by the National Bureau of Statistics in 2018, we will perform the weighted sampling of all patients proportionally from the seven regions, as shown in table 2.22 Table 1Estimated sample size for primary study objective considering different precision levelsType of cancerEstimated number of incident cases per year in ChinaEstimated HPV-AF (%)Sample size for specific precision levelsSample size5%3%Oropharynx6018*22.83*271–271Non-oropharynx110 787550Oral cavity20 831*4.30*–176176Nasopharynx60 558†5.90‡86–86Hypopharynx5990†2.40†–100100Larynx23 408*4.60*–188188Total116 805821*From Duan et al.15†From Bray et al.31‡From Castellsagué et al.32AF, attribution fraction; HPV, human papillomavirus. Estimated sample size for primary study objective considering different precision levels *From Duan et al.15 †From Bray et al.31 ‡From Castellsagué et al.32 AF, attribution fraction; HPV, human papillomavirus. Table 2Sample distribution by population density and geographic regionGeographic regionsNumber* (10 000)Population proportion (%)Sample distribution by year periods*Total2008–20092013–20142018–2019Northeast10 8367.8204686868North17 52212.5324108108108Northwest10 2797.4192646464East36 47726.1678226226226Central27 06919.4504168168168South17 20612.3321107107107Southwest20 21714.5378126126126Total139 6061002601867†867†867†*Roughly sample allocation.†The sample size increase considers 5% of tissue samples will not be valid for evaluation. Sample distribution by population density and geographic region *Roughly sample allocation. †The sample size increase considers 5% of tissue samples will not be valid for evaluation. Site selection and patient recruitment Two tertiary hospitals are selected from each of the seven geographic regions across China (Northern, Northwestern, Northeastern, Central, Southern, Southwestern and Southeastern), considering cultural and socioeconomic diversity in China (figure 2). All selected hospitals are cancer centres which serve as central referral treatment facilities and treat the majority of patients with newly diagnosed HNC from the surrounding cities, towns and villages. Hence, the participants in BROADEN-China are believed to broadly represent patients with HNC in Chinese populations. The study protocol and related documentation were approved by the institutional review board (IRB) of the leading site—Henan Cancer Hospital (2020-KY-0008-001) and other participating hospitals. A waiver for patient informed consent was approved by IRB for all sites. Participating hospitals will examine their archived medical records from all patients with a first primary diagnosis of HNC with pretreatment formalin-fixed paraffin-embedded (FFPE) tissue samples stored for the time periods 2008–2009, 2013–2014 and 2018–2019. Eligible patients will be consecutively enrolled until the sample size has been reached for each study period by a specific anatomic site. Map of geographic regions in China and estimated sample size per region, considering the three time periods. Sample sectioning, histological assessment and HPV testing After eligible patients are selected, each site will be asked to provide the sample blocks available for sectioning. Well-trained technicians from the central lab will travel to each site to perform the block sectioning. At least seven paraffin sections will be obtained from each block using the sandwich method. Considering the unique specimen restriction in China, block sectioning, H&E and p16INK4a IHC staining, and EBER-ISH testing (only for NPC) will be performed at each participating site. On receiving the tissue sections at the central laboratory, paraffin block processing and pathology reassessment of the initial histopathological diagnosis (including histological subtype and stage) will be conducted by pathologists at the central lab using the consensus criteria established by an expert panel of pathologists based on the WHO classification of HNCs. Five per cent of samples will be rediagnosed by the expert panel for quality control. The discrepancy will be resolved by the pathologist from the global central lab. On passing the pathology reassessment, HPV DNA will be extracted by Maxwell 48 automated nucleic acid station combined with the compatible kits. The extracted DNA will be then amplified by consensus primer SPF10 and detected by DEIA, followed by LiPA25 genotyping. Meanwhile, p16INK4a immunohistochemistry will be performed on all samples as a surrogate biomarker for the presence of HPV. All ‘HPV-DNA-positive’ and ‘HPV-DNA-negative’ but ‘p16INK4a positive’ samples will undergo RNA extraction by PureLink FFPE Total RNA Isolation Kit (Invitrogen), E6*I mRNA amplification by QuantiTect Virus Kit (Qiagen) and detection by Luminex 200. A random selection of 10% of the remaining ‘HPV-DNA-negative’ samples will also be tested for HPV 16 E6*I mRNA. Ten percent of all HNC samples will be retested for quality control, in which discrepancy is resolved by a third test. Data collection and analysis Our study uses two types of medical records, paper-based records in 2008–2009 and electronic records in 2013–2014 and 2018–2019. Both records will be transcribed into the database by two independent researchers to ensure the accuracy and completeness of all records. The data captured in the study database will be used for all analytic endpoints. In addition to the variables collected in the BROADEN study in other countries (patient demographics, primary tumour characteristics and risk factors for HNC including smoking status, alcohol consumption and sexually transmitted infections),20 China-specific risk factors such as betel nut chewing and EBV status for nasopharynx cancers will be documented. Logistic regression models will be performed to assess the differences in HPV-AFs between three time periods (independent variables) adjusting for all covariates. Finally, the data of BROADEN-China in 2008–2009 and 2018–2019 will be used for a joint analysis with other countries in the BROADEN-Global study. Patient and public involvement The case report form has been pilot-tested in the leading site Henan Cancer Hospital. Discussion BROADEN-China will be the first and largest cross-sectional study estimating the HPV-AFs of HNC in China with robust clinical practice and lab testing techniques. It is one part of BROADEN-Global, with tailored design and implementation adaptations to reflect China-specific settings. These adaptations include the addition of a time point (2013–2014), the inclusion of China-specific risk factors for HNCs (betel nut chewing and EBV) and the transformation and validation of the paper-based medical records to our database. Also, samples recruited based on the population density from regions ensured the generalisability of the study results to a national level. In summary, BROADEN-China will establish robust protocols for implementing large national, multisite, cross-sectional epidemiological studies in China to explore the incidence or time trend of cancer and other diseases. BROADEN-China will not only contribute to global knowledge of HPV-AFs on OPC, but also NPC, especially in Southern China where NPC is more prevalent compared with Europe and the USA.21 23 In China, NPC is characterised by distinct gender, ethnic and geographical distribution, with 2.5 times greater prevalence in males than in females, and substantially higher rates observed among the Cantonese population in Southern China.24 The presence of EBV in NPC might interfere with the HPV attribution because of a potential interaction between EBV and HPV.25–27 Importantly, to the best of our knowledge, there is no national data to report the HPV-AFs on NPC in China. BROADEN-China will describe the impact of HPV-AFs on NPC, which could also be taken as a reference in indigenous populations in Southeast Asia, the Arctic region, North Africa and the Middle East,28 where NPC is also prevalent. Compared with the other mainland China-based studies examining HNC burdens in China,11–13 17 19 BROADEN-China has additional merits. This is the first HPV attributability study in China using the combination of robust testing of HPV DNA, HPV E6*I mRNA and p16INK4a, which will allow the study to minimise bias associated with sole-HPV testing techniques. When conducting PCR-based testing for FFPE specimens, parts of the viral genome can be fragmented during the long periods of storage or the process of tissue sectioning.29 The sequence length of the PCR primer becomes essential to ensure high HPV DNA test sensitivity for FFPE samples. The SPF10-DEIA-LiPA25 assay amplifies a 65 bp fragment within L1.30 The small size of the amplicon makes SPF10-DEIA-LiPA25 particularly useful for testing of FFPE tissue blocks, compared with other assays. Second, our sampling strategy ensures diversity of the population to ensure the national representativeness. According to economic and cultural differences in China, two representative tertiary or cancer hospitals are selected in each geographic region. Due to the difficulties in diagnosis and multimodality treatment approaches of HNC, patients usually have to visit local tertiary hospitals or tumour-specialised hospitals for diagnosis and treatment, which increases the representativeness of our study. Lastly, this study will report the HPV-AFs across three different time periods. Estimation of the current and evolving contribution of HPV infection to HNC is important to explore the disease burden of HNC and to inform policymakers of the potential impact on regional and national populations of primary prevention interventions on HNCs. Several limitations should also be considered. This study includes pretreatment biopsy samples stored at the pathology laboratory and qualified for testing. This may be of particular concern for the earlier time period assessed (2008–2009), where the limited data quality and huge missing data in the paper-based records may impact the power and precision of the HPV-AF estimation. Additionally, due to the lack of China-specific data for the incidence rate of HNC in China, population density was used as an alternative approach to determining sample sizes with weighting instead of the weighted number of HNC cases in each region. BROADEN-China will provide a more comprehensive understanding of the historical burden and changes of HNC attributable to HPV infection across the study periods in China. BROADEN-China will support the development of future protocols for implementing large national, multisite, cross-sectional epidemiological studies in China to explore incidence or time trends of diseases (especially for cancers). Supplementary Material Reviewer comments Author's manuscript