Discussion

It is known that head and neck cancers are one of the most common malignancies, and they contribute to around 300 000 deaths each year.^{Reference Jemal, Bray, Center, Ferlay, Ward and Forman1} As mentioned above, 90 per cent of the tumours are squamous in origin. This is also reflected in head and neck cancer research as more than 80 per cent of the top 100 cited articles focused on SCCs. A more in-depth bibliometric analysis of head and neck cancer research to date offers the reader some interesting insights into this field of study, its history and how treatment for head and neck cancer has evolved over the years.

The most cited paper on the list was published in The New England Journal of Medicine by Bonner et al.^{Reference Bonner, Harari, Giralt, Azarnia, Shin and Cohen10} in 2006. This paper described a randomised trial comparing radiotherapy alone with radiotherapy plus cetuximab in the treatment of locoregionally advanced head and neck SCC. Their results showed that radiotherapy plus cetuximab improves locoregional control and reduces mortality, without increasing common radiotherapy-related toxic effects.

The second most cited article was by Ang et al.^{Reference Ang, Harris, Wheeler, Weber, Rosenthal and Nguyen-Tan12} in 2010 in the The New England Journal of Medicine. In their retrospective analysis, the authors investigated the association between tumour HPV status and survival among patients with stage III or IV oropharyngeal SCC. The study reported a higher survival rate and lower risk of death in HPV-positive tumours. This landmark paper also provided strong evidence for HPV testing in oropharyngeal SCCs, which contributed to current routine practice and the use of cyclin-dependent kinase inhibitor p16 as a prognostic biomarker.

The third most cited paper was published by Gillison et al.^{Reference Gillison, Koch, Capone, Spafford, Westra and Wu13} in 2000, in the Journal of the National Cancer Institute. The authors retrospectively evaluated the relationship between HPV and head and neck cancer, focusing on the aetiological role of HPV in head and neck SCCs. They concluded that HPV-positive oropharyngeal cancers are causally associated with HPV infection and comprise a distinct disease entity.

When considering citation numbers of the top 100 cited articles, there is a significant discrepancy between the citation numbers of the top two cited papers (3205 and 2731) and the rest of the papers on the list (range, 410–1831). The number of citations for the most cited paper was approximately 4 times that of the mean citation number of the list (828) and 5 times that of the median (626). These trends may reflect the attention that clinicians and researchers have on the effectiveness and outcomes of treatment with cetuximab, an epidermal growth factor receptor inhibitor, in head and neck cancers. This is shown by the fact that several other articles on the top 100 list also involved cetuximab as part of their studies.^{Reference Bonner, Harari, Giralt, Cohen, Jones and Sur16–Reference Vermorken, Trigo, Hitt, Koralewski, Diaz-Rubio and Rolland18}

To evaluate the quality and influence of the top 100 cited papers, a review of the journals represented by the top 100 articles list and their respective journal impact factors could be performed. Journal impact factors are formed on the basis of citations from within the scientific community, as they are the quantification of the average citations received by articles published in a journal during a specified period. More influential papers are more likely to be cited. Hence journals with higher impact factors are likely to contain more influential articles. This idea is supported by the fact that 8 of the top 10 most cited papers were from 2 journals with very high impact factors of more than 50 (The New England Journal of Medicine and The Lancet) and that these two journals contributed nearly one-fourth of the papers on the list. As expected, this shows that high impact factor journals heavily influence head and neck cancer research through their high-quality publications. Nevertheless, it would be of value for future research to identify the association between citation numbers and clinical influence, by determining whether papers with higher citation numbers are more regularly incorporated into clinical practice.

The oldest article on the list was published by Henle et al.^{Reference Henle, Henle, Ho, Burtin, Cachin and Clifford14} in 1970, in the Journal of the National Cancer Institute. The authors titrated serum samples that were prospectively collected from head and neck cancer patients and control groups from several countries, to look for antibodies to Epstein–Barr virus. The data confirmed an unusually high frequency of anti-Epstein–Barr virus titres among patients with nasopharyngeal carcinoma. On the other hand, the most recent paper was a randomised phase 3 trial published by Ferris et al.^{Reference Ferris, Blumenschein, Fayette, Guigay, Colevas and Licitra15} in The New England Journal of Medicine in 2016. This study compared treatment outcomes of nivolumab, an anti-programmed death 1 monoclonal antibody, with standard single-agent therapy for patients with recurrent or metastatic head and neck SCCs after platinum chemotherapy. They concluded that nivolumab leads to prolonged overall survival for these patients.

The comparison of different treatment modalities for head and neck cancers is also a common theme observed when analysing the recurrent topics represented within the list. Almost a third of the papers on the list compared different treatment methods, with chemotherapy being the most heavily studied subject. Another recurrent theme is the relationship between HPV and head and neck cancers, with 16 of the 100 papers published focusing on this area. Most of these papers focused on the association between HPV and different subgroups of head and neck cancers, or the prognosis of HPV-positive head and neck cancers.

Examining the manuscripts on the list by decade shows some interesting trends in the progression of head and neck cancer research. Papers from the 1990s focused on the comparison of different treatment modalities for head and neck cancers. Several randomised trials from this decade compared treatment outcomes of radiotherapy alone versus concomitant radiotherapy and chemotherapy^{Reference Brizel, Albers, Fisher, Scher, Richtsmeier and Hars19–Reference Wendt, Grabenbauer, Rodel, Thiel, Aydin and Rohloff21} in head and neck SCCs, highlighting the importance of chemotherapy in the treatment of head and neck cancers. A few laboratory studies looking into the basic science and genetics of head and neck cancers were also published during this period.^{Reference Califano, van der Riet, Westra, Nawroz, Clayman and Piantadosi22–Reference Schipper, Frixen, Behrens, Unger, Jahnke and Birchmeier24} Randomised studies comparing different treatment methods remained an essential focus among papers published in the 2000s. Instead of merely focusing on the widely available traditional treatments, more papers looked into the development of novel therapeutic agents, such as selective epidermal growth factor receptor tyrosine kinase inhibitors.^{Reference Cohen, Rosen, Stadler, Recant, Stenson and Huo25,Reference Soulieres, Senzer, Vokes, Hidalgo, Agarwala and Siu26} It is also during this period that articles investigating the relationship between HPV and head and neck cancers started to appear on the list,^{Reference Gillison, Koch, Capone, Spafford, Westra and Wu13,Reference Herrero, Castellsague, Pawlita, Lissowska, Kee and Balaram27,Reference Mork, Lie, Glattre, Hallmans, Jellum and Koskela28} with some of them reporting a better prognosis in patients with HPV-positive head and neck SCCs.^{Reference Ragin and Taioli11,Reference Fakhry, Westra, Li, Cmelak, Ridge and Pinto29} A similar pattern is seen for manuscripts published in the 2010s, as a number of them also examined the roles of potential new drugs^{Reference Ferris, Blumenschein, Fayette, Guigay, Colevas and Licitra15,Reference Seiwert, Burtness, Mehra, Weiss, Berger and Eder30} and the relationship between HPV and the rising incidences of oral cavity and oropharyngeal cancers.^{Reference Chaturvedi, Engels, Pfeiffer, Hernandez, Xiao and Kim3,Reference Chaturvedi, Anderson, Lortet-Tieulent, Curado, Ferlay and Franceschi31}

An analysis of the papers according to the country of origin was also performed. As mentioned above, the USA contributed more than 65 per cent of the top 100 articles in head and neck cancer. This phenomenon is also seen in other bibliometric analyses in otolaryngology.^{Reference Fenton, Roy, Hughes and Jones4,Reference Saunders, Rymer and McNamara5} Ellul et al.^{Reference Ellul, Bullock, Abdelrahman, Powell, Witherspoon and Lewis7} attributed this trend to the preference of USA institutions to cite ‘local research’ in their citation analysis on emergency abdominal surgery. They pointed out that the USA has an influential research culture in medical training, which encourages the integration of research with clinical practice. Moreover, in a large-scale bibliometric study of prospective head and neck SCC therapeutic research that investigated the effects of study sponsorship on the publication in this field,^{Reference Sun, Houlton, Moloci, MacEachern, Bradford and Prince32} it is found that the USA produced the highest number of publications and that the National Institutes of Health was the second largest sponsor of the research overall.

There are several limitations to our study. Some articles might have received multiple or disproportionate citations due to institutional bias, language bias, powerful-person bias or self-citation. This effect might also have been fortified by limiting results to English language papers during the article search process. Moreover, citation numbers are often influenced by the age of paper. Having been in the public domain for a more extended period, older papers might have accumulated a higher number of citations. As for more recent influential papers, more time is needed to accumulate citations due to publication lead-time for their citing manuscript. However, more than 70 of the top 100 cited papers on our list were from 2000 onwards, with the most recent article published in 2016. This confirms the importance and influence of the more recent articles and suggests that publication time bias might not be as significant in the field of head and neck cancer research.

In summary, this article provides an overview of some of the most influential papers in the study of head and neck cancers despite the above flaws of bibliometric analyses. It offers some insight into the progression of our understanding of head and neck cancers as well as treatment and related research. For those who currently contribute to the development of the literature surrounding head and neck cancer, this paper highlighted some essential patterns and objectives represented in the literature and thus serves as an indicator and guide for them. This analysis could inform future research priorities as well as stimulate discussion within the scientific community.