Global research on sinonasal inverted papilloma over the past two decades: a bibliometric analysis| ACTA Otorhinolaryngologica Italica

Abstract

Objective. This study aimed to investigate the global research status, hot topics, and prospects in the field of sinonasal inverted papilloma (SNIP) through bibliometric analysis.
Methods. The literature on SNIP was retrieved and downloaded from the Web of Science Core Collection from 2002 to 2021. The bibliometric and visualisation networks of SNIP were constructed using VOSviewer 1.6.18, CiteSpace 6.1. R2, and a bibliometric online analysis platform.
Results. A total of 560 original articles about SNIP research were included, involving 2,457 authors from 610 institutions in 45 countries. The number of SNIP publications showed an overall rising trend, with an average annual output of 28 articles and almost 3 times as many articles published in 2020 as in 2002. The analysis of keyword burst detection indicated that EGFR mutation, malignant transformation and infection are emerging research hotspots. Moreover, EGFR mutation, KRAS mutation, malignant tumour, metallothionein 2a gene, pre-operative diagnosis, HPV-negative tumour, and expression were among the 11 key clusters of co-cited references.
Conclusions. This study provided a comprehensive, systematic, and objective analysis and visualised knowledge map of SNIP over the past 2 decades. In particular, current hotspots and prospective trends in the field of SNIP have been identified. These results highlight the future direction of SNIP research for rhinologists.

Introduction

Sinonasal inverted papilloma (SNIP) is a benign epithelial neoplasm that commonly occurs in the nasal cavity or paranasal sinuses, initially described by Ward in 1854 ¹. SNIP is histologically characterised by endophytic growth of the surface respiratory epithelium into underlying connective tissue stroma without involving the basement membrane ². The occurrence of SNIP can lead to recurrence and carries a risk of transformation into squamous cell carcinoma (SCC). The primary causes of SNIP recurrence are the surgeon’s experience, tumour location and inadequate removal ³. It is essential to note that SNIP has a rate of 7.1% for synchronous carcinoma and in 3.6% of cases develops into metachronous carcinoma ⁴. Currently, no specific features have been identified as significant predictors of multiple recurrences or malignancy. In general, radical surgery with post-operative radiation is the preferred treatment for malignant SNIP. There are few studies that provide a comprehensive overview of research trends in SNIP, including its aetiology, diagnosis and treatment strategies.

Bibliometric analysis serves as a crucial statistical method to analyse academic information and scientific outcomes of the medical literature on a selected research area over a given period using computational and visual analytic approaches. Unlike traditional systematic reviews and meta-analyses, bibliometric analysis can reveal both qualitative and quantitative aspects of a research topic from a systematic and intuitive level, exploring its longitudinal and transversal characteristics, developments, and multiple ramifications across various research fields ⁵. This analysis has been widely utilised in various medical specialties, including cerebrospinal fluid leak, allergic rhinitis, and chronic rhinosinusitis, to gain a deeper understanding of pathogenesis, identify therapeutic alternatives, improve prognosis and develop guidelines ^6-8. However, bibliometric analysis related to SNIP has not been extensively conducted.

In this study, we aimed to systematically compile the available literature on SNIP from 2002 to 2021, comprehensively create collaboration networks and keyword clusters, and reveal emerging trends and hot spots to advance future research and clinical applications in the forthcoming years. To this end, we constructed the organic image of SNIP to portray an overview from a historical standpoint, thereby providing a new perspective for rhinologists to facilitate the formulation and administration of scientific research.

Materials and methods

Data source and search strategy

The bibliometric study was carried out using the Science Citation Index Expanded (SCI-E) and the Social Sciences Citation Index (SSCI) of the Web of Science Core Collection (WoSCC). We performed a thorough search of the WoSCC database for relevant publications from 2002 to 2021 and included only original articles. The search terms that we used in the WoSCC were as follows: Topic = (“nasal inverted papilloma” OR “sinonasal inverted papilloma”) AND Language = English. To avoid possible bias from frequent updated databases, all data downloads and systematic literature searches were documented and completed on October 1, 2021. The detailed search procedure is shown in Figure 1A. Two researchers independently undertook data analysis. Any divergences were resolved through negotiation or by seeking the help of a third reviewer to reach an excellent consensus. The full records of data were documented and screened for inclusion based on titles, journals, abstracts, keywords, authors, countries/regions, institutions, references, Hirsch index (H-index) in 2021 and total citations. The H-index is considered an important indicator of the scientific impact of an author’s publications.

Analysis tool

WoSCC data were converted to text before being imported into the software. VOSviewer 1.6.18 (Leiden University, Leiden, the Netherlands), CiteSpace 6.1. R2, 64-bit (Drexel University, Philadelphia, PA, USA) and a bibliometric online analysis platform () were used to analyse co-cited articles, keywords, countries, institutions, journals, authors and network features of keyword bursts, as well as to present the results visually.

Exported records were in “plain text” format with “full records and citation references”, and were imported into CiteSpace. CiteSpace, a bibliometric analysis visualisation software, was used to investigate institutional collaboration networks, detect co-cited references, explore keywords and references with high citation bursts over time, and forecast the research hotspots and emerging trends in this study. The parameters of CiteSpace were set as follows: time slicing from 2002 to 2021, years per slice (1), term source (all selection), node type (choose one at a time), links (strength: cosine; scope: within slices), selection criteria (the scale factor κ = 25), pruning (pathfinder, pruning sliced networks, pruning the merged network), and visualisation (cluster view-static, show merged network). Additionally, the centrality score was analysed using CiteSpace. The centrality score assesses the relevance of network nodes, with the greater centrality being a more important node. On the network visualisation map, the size of the nodes represents the frequency, whereas the different colours of the nodes represent different years. The lines between nodes imply cooperation or co-occurrence or co-referencing relationships.

Similarly, VOSviewer was employed to visually illustrate the co-occurrence networks of authors and keywords, classify keywords into different clusters, and construct density maps. Cluster analysis of research hotspots can be visualised, and the keyword co-occurrence network can identify emerging trends.

Results

Annual outputs and growth trends

In total, 560 original articles about SNIP were published from 2002 to 2021. Despite some fluctuations between 2012 to 2016, research outputs on SNIP exhibited an overall growing trend from 2002 to 2021, with almost three times as many articles published in 2020 as in 2002 (Fig. 1B). The annual average number of publications was 28. The number of publications per year has exceeded 20 since 2006. The number of publications reached a peak of 47 in 2020, exceeding 40 for the first time. A growth trend model (R2 = 0.7716) indicated that the volume of literature on SNIP would continue to increase and show potential in the near future.

Distribution of countries/regions and institutions

The publications in the field were distributed among 610 institutions from 45 countries. Twenty-four countries and 573 institutions each contributed at least five articles. The top 10 contributing countries are listed in Table I. Most publications came from the United States (n = 139), followed by China (n = 88), Japan (n = 53), South Korea (n = 45) and Italy (n = 44). The United States was the most productive country, publishing > 24% of the studies. The centrality score indicates the relative importance of nodes in a network. Higher centrality in a collaborative network determines more frequent cooperation. The United States (0.28) had the highest centrality, closely followed by Italy (0.27), China (0.16) and Canada (0.16). Table II revealed the top 10 institutions that published the highest number of articles on SNIP from 2002 to 2021, including Capital Medical University (n = 26), Fudan University (n = 15), and Sun Yat-sen University (n = 9) in China. The universities with the highest centrality score were the University of Pennsylvania (0.09) in the United States, followed by South Korea’s Pusan National University (0.05).

Contributions of authors

We found that 2457 authors published articles on SNIP from 2002 to 2021. Table III lists the top 10 most productive authors in terms of the number of published papers. Luo Zhang and Chengshuo Wang of Capital Medical University (13 publications with 142 citations; 13 publications with 135 citations, respectively) had the most published manuscripts, followed by James N. Palmer of the University of Pennsylvania (11 publications; 210 citations) with an H-index of 48. Dehui Wang of Fudan University ranked first in terms of the centrality score (0.02) with an H-index of 25. The author with the highest H-index of 68 was David W. Kennedy, followed by James N. Palmer.

Cluster analysis of keyword co-occurrence related to research hotspots

In total, 1,598 different relevant keywords were extracted from 560 articles. A network map was generated for keywords with more than 5 occurrences, which includes 184 terms in the map (Fig. 2A). In the map, the high-frequency keywords were “inverted papilloma” (n = 294), “paranasal sinuses” (n = 149), “management” (n = 146), “nasal cavity” (n = 99), “surgery” (n = 98), “recurrence” (n = 74), “head” (n = 73), “sinonasal inverted papilloma” (n = 72), “nose” (n = 63), and “expression” (n = 62). The circle size represented the number of keyword occurrences. Terms associated with similar research topics were grouped into the same catalogue. The keywords were classified into five main clusters: the purple cluster was related to the anatomy of the nasal cavity and sinuses, the red cluster was associated with the basic research of malignant transformation, the blue cluster was linked to clinical diagnosis, the green cluster pertained to surgical treatment, and the yellow cluster was primarily associated with the risk factors for recurrence (Fig. 2A).

We next coded different keywords by colour in order of time in the corresponding literature (Fig. 2B). The keywords in blue appeared earlier, and the keywords in yellow appeared later. The early research mainly focused on the themes of diagnosis and gradually developed into surgical treatment. Yellow-green colours represent the most recent research topics, and the representative keywords were “recurrence”, “classification”, “risk factor”, “EGFR mutation” and “malignant transformation”. The results showed recurrence and malignant transformation might become future research emphases.

Detection of keyword bursts

The keywords with the strongest citation bursts can represent the previous or current research hotspots, predict the future development trends of the field, and highlight potential hotspots in a discipline. The chronology is displayed by a blue timeline that spans across years. A red segment denotes the start year, end year, and time span of a burst. Keyword bursts between 2002 and 2021 were detected based on an examination of 560 articles included in the WoSCC database (Fig. 2C). In the early stage, “nose” had the strongest burst (strength = 12.89). “EGFR mutation” has emerged in the last three years (2018-2021) with a strength of 6.64, followed by “recurrence” (strength = 5.93) and “meta-analysis” (strength = 3.88). Notably, the bursts of “EGFR mutation”, “malignant transformation” and “infection” are still ongoing.

Analysis of co-cited references

From 560 articles, 7797 cited references were submitted for analysis of co-citation correlation, and a cluster network diagram was created based on the results. The nodes of the network represent references. The literature co-citation network consisted of 80 nodes and 146 links (Fig. 3A). The node size is proportional to the total number of co-citations of the article. The node colour represents the cited time of reference. The nodes with a thick purple ring around them (Cannady ⁹ and Mirza ⁴) were used to connect the stages of the growth of a field, which implied to be the turning points and hubs of a domain and play an important undertaking role in the entire internal network. The red rings (Wang ¹⁰ and Udager ¹¹) indicated to be explosive citations. By identifying the crucial keywords from the cited references, we could detect the research hotspots. “EGFR mutation”, “KRAS mutation”, “malignant tumour”, “metallothionein 2a gene”, “pre-operative diagnosis”, “HPV-negative tumour”, and “expression” were among the 11 key clusters of co-cited references (Fig. 3B).

Discussion

This is the first application of quantitative and qualitative bibliometric methods to SNIP research. It involved 560 research papers retrieved from WoSCC database. Over the past 20 years, nearly a quarter of all articles were from the United States, which greatly promoted the advancement of SNIP research. In addition, the University of Pennsylvania had the strongest partnerships with other institutions, which were beneficial for those institutions that rarely communicate with each other. The United States demonstrated the most intensive collaboration and the highest centrality to other countries. Moreover, an increasing number of scholars worldwide showed interest in SNIP research according to the growth curve. This trend partly implied a mature medical research and health environment in these countries and the critical need for efficient SNIP research. South Korea, France, United Kingdom, and Poland also published many papers. However, they may have less collaboration with other countries, as shown by their relatively low centrality scores. The countries with lower rates of cooperation should strengthen their international collaborations, especially with those leading nations, in order to accelerate their progress in SNIP research.

Herein, we focused on further detection of the co-occurrence and burst analysis of major keywords. We performed generation of clustering based on co-occurrence information of main keywords. According to similar research subjects under the same catalogue, we grouped them into five categories: anatomy, pathogenesis, diagnosis, treatment, and recurrence of SNIP. Based on keyword bursts in the last five years, we discussed the relationship between classification staging systems and recurrence of SNIP. At the same time, most current studies on malignant transformation of SNIP have expressed concerns about the research subjects and quality, and so it is meaningful to make a general summary based on them. Based on the above discussion of two clusters, we gained a profound understanding of present concerns and research hotspots in this field.

Although endoscopic sinus surgery has been considered a treatment of choice for most SNIP cases, better treatment planning should account for the clinical stage which is closely associated with post-operative outcomes. In 2000, Krouse first proposed a staging system that incorporated both the extent and volume of tumour, which has been widely adopted over the last few decades ¹². However, several studies found that recurrence rates did not differ significantly between different stages of the Krouse staging system ^13,14. Subsequently, other grading systems have been improved over time ^9,15-17. Nakayama and colleagues revealed that Meng et al.’s origin site-based staging system maintained a proportional hazard ratio for all follow-up durations without crossing the survival curve and had a good correlation with recurrence rate ¹⁸. Meng et al. also recommended optimal surgical strategies according to their staging system and demonstrated a total recurrence rate of 6.4% involving 675 SNIP patients in a subsequent prospective multicentre study, which was much lower compared with several earlier large sample studies ¹⁷. Clearly, it is very important to accurately predict where tumours originate before surgery. The osteitis/neosteogenesis signs on computed tomography combined with convoluted cerebriform patterns reverse-tracing method on T2- or enhanced T1-weighted images have been employed to predict and identify original site of SNIP ^19,20. The explanations for underlying pathological mechanisms involved the different distributions of cellularity, vascularity and claudin-5 expression in microvessels between original site and periphery of SNIP ²¹.

Despite its tendency for malignant transformation, predicting which patients with SNIP will transform into SCC is difficult as there is no correlation between clinical variables, such as the number of SNIP local recurrences or surgeries, and the subsequent progression to carcinoma ²². Different molecular mechanisms are suspected to play crucial roles in this process. Previously, SNIP-associated SCC had significantly lower rates of positive margins, advanced T classification and distant metastasis compared to de novo SCC ^23,24. These findings suggested the involvement of distinct molecular pathways in oncogenesis. Several studies revealed that there seemed to be a strong association between infection with high-risk HPV subtypes and malignant transformation of SNIP ^25-27. HPV-18 showed the greatest overall average effect size of common high-risk subtypes ^26,27. Moreover, EGFR mutations have also been implicated in the pathogenesis of SNIP-associated SCC, but their relationship to HPV status is unclear. Identical EGFR genotypes were found in matched pairs of SNIP and SNIP-associated SCC. Treatment of SNIP-associated carcinoma cells with irreversible EGFR inhibitors resulted in the inactivation of EGFR signaling and growth inhibition ²². Conversely, EGFR mutations characterised SNIP with a lower risk of developing into SCC ²⁸. As a result, further research into the risk of malignant progression in relation to SNIP is required.

Limitations

This study had several limitations in using bibliometric analysis for research on SNIP. First, we only used data from the SCI-E and SSCI of WoSCC between 2002 and 2021, and did not include data from other relevant search engines and other time periods. Some publications from other databases might have been missed. However, the WoSCC database is one of the most extensive and well-recognised global resources and has been extensively applied for the retrieval of natural science publications. The overall situation and general trend in SNIP treatment over the past two decades were fully demonstrated. Second, the consideration of articles published in English only may have generated a linguistic bias in our analyses, even if English is the most commonly used language to publish academic documents worldwide. Finally, the omission of books/chapters/letters may not fully represent all publications about SNIP. However, original articles are the dominant types of publications and well elucidate the status and progress of disease research.

Conclusions

In this study, we conducted a bibliometric analysis between 2002 and 2021 and provided an overview of SNIP research. Over the past two decades, the United States accounted for more than 24% of the global publications on SNIP, which have made significant contributions to the field. Notably, the number of articles in the field of SNIP research worldwide has increased annually. We analysed the developments and trends of major keywords such as recurrence and malignant transformation, which can help rhinologists discover new insights for further research and provide valuable guidance for topic selection.

Conflict of interest statement

The authors declare no conflict of interest.

Funding

This work was supported by Natural Science Foundation of Anhui Provincial Education Department (Grant Number: KJ2021A0315).

Author contributions

DF designed the study, collected the data, and drafted the manuscript. YZ analysed the data. JY provided instruction on the use of relevant analytical applications and supervised the study. JH re-examined and analysed the data. All authors critically reviewed and revised the manuscript. All authors approved the submitted version.

Ethical consideration

This study was approved by the Institutional Ethics Committee of the Second Affiliated Hospital of Anhui Medical University (approval number/protocol number YX2021-036).

The research was conducted ethically, with all study procedures being performed in accordance with the requirements of the World Medical Association’s Declaration of Helsinki.

Written informed consent was obtained from each participant/patient for study participation and data publication.

Data availability statement

The raw datasets supporting this study’s findings can be found in the article/supplementary material online. Further inquiries can be directed to the corresponding authors.

Figures and tables

Figure 1.Frame flow diagram of search strategy (A) and trends in the number of publications (B) of SNIP research from 2002 to 2021.

Figure 2.Co-occurrence analysis of global research on SNIP based on the WoSCC database from 2002 to 2021. (A) Mapping of keywords in the research field. (B) Distribution of keywords according to the chronological order of appearance. (C) Keywords with the strongest citation bursts in SNIP research.

Figure 3.Co-cited references map (A) and clustered map of co-cited references (B) on SNIP research from 2002 to 2021.

References

1. Ward N. A mirror of the practice of medicine and surgery in the hospitals of London: London Hospital. Lancet. 1854; 2:480-482.
2. Ringertz N. Pathology of malignant tumors arising in the nasal and paranasal cavities and maxilla. Acta Otolaryngol. 1938; 27:31-42.
3. Goudakos JK, Blioskas S, Nikolaou A. Endoscopic resection of sinonasal inverted papilloma: systematic review and meta-analysis. Am J Rhinol Allergy. 2018; 32:167-174. DOI
4. Mirza S, Bradley PJ, Acharya A. Sinonasal inverted papillomas: recurrence, and synchronous and metachronous malignancy. J Laryngol Otol. 2007; 121:857-864. DOI
5. You Y, Wang D, Wang Y. A bird’s-eye view of exercise intervention in treating depression among teenagers in the last 20 years: a bibliometric study and visualization analysis. Front Psychiatry. 2021; 12:661108. DOI
6. Ramos-Rincon JM, Monjas-Canovas I, Abarca-Olivas J. Research in cerebrospinal fluid leak (rhinorrhea and otorrhea): a bibliometric analysis from 1945 to 2018. Cureus. 2022; 14:e21888. DOI
7. Wu Q, Zheng R, Wang W. The top 100 most influential articles in allergic rhinitis from 1970 to 2018: a bibliometric analysis. J Int Med Res. 2019; 47:6315-6336. DOI
8. Zhou F, Zhang T, Jin Y. Developments and emerging trends in the global treatment of chronic rhinosinusitis from 2001 to 2020: a systematic bibliometric analysis. Front Surg. 2022; 9:851923. DOI
9. Cannady SB, Batra PS, Sautter NB. New staging system for sinonasal inverted papilloma in the endoscopic era. Laryngoscope. 2007; 117:1283-1287. DOI
10. Wang MJ, Noel JE. Etiology of sinonasal inverted papilloma: a narrative review. World J Otorhinolaryngol Head Neck Surg. 2017; 3:54-58. DOI
11. Udager AM, McHugh JB, Goudsmit CM. Human papillomavirus (HPV) and somatic EGFR mutations are essential, mutually exclusive oncogenic mechanisms for inverted sinonasal papillomas and associated sinonasal squamous cell carcinomas. Ann Oncol. 2018; 29:466-471. DOI
12. Krouse JH. Development of a staging system for inverted papilloma. Laryngoscope. 2000; 110:965-968. DOI
13. Kim JS, Kwon SH. Recurrence of sinonasal inverted papilloma following surgical approach: a meta-analysis. Laryngoscope. 2017; 127:52-58. DOI
14. Lisan Q, Moya-Plana A, Bonfils P. Association of Krouse classification for sinonasal inverted papilloma with recurrence: a systematic review and meta-analysis. JAMA Otolaryngol Head Neck Surg. 2017; 143:1104-1110. DOI
15. Oikawa K, Furuta Y, Nakamaru Y. Preoperative staging and surgical approaches for sinonasal inverted papilloma. Ann Otol Rhinol Laryngol. 2007; 116:674-680. DOI
16. Han JK, Smith TL, Loehrl T. An evolution in the management of sinonasal inverting papilloma. Laryngoscope. 2001; 111:1395-1400. DOI
17. Meng Y, Fang G, Wang X. Origin site-based staging system of sinonasal inverted papilloma for application to endoscopic sinus surgery. Head Neck. 2019; 41:440-447. DOI
18. Nakayama T, Tsunemi Y, Kashiwagi T. Comparison of current staging systems for sinonasal inverted papilloma. Am J Rhinol Allergy. 2021; 35:64-71. DOI
19. Nakamaru Y, Fujima N, Takagi D. Prediction of the attachment site of sinonasal inverted papillomas by preoperative imaging. Ann Otol Rhinol Laryngol. 2014; 123:468-474. DOI
20. Fang G, Lou H, Yu W. Prediction of the originating site of sinonasal inverted papilloma by preoperative magnetic resonance imaging and computed tomography. Int Forum Allergy Rhinol. 2016; 6:1221-1228. DOI
21. Ma S, Xian M, Yang B. Pathological changes from the originating to the peripheral sites of sinonasal inverted papilloma are the underlying mechanisms of preoperative MRI-tumor origin prediction. Rhinology. 2020; 58:59-65. DOI
22. Udager AM, Rolland DCM, McHugh JB. High-frequency targetable EGFR mutations in sinonasal squamous cell carcinomas arising from inverted sinonasal papilloma. Cancer Res. 2015; 75:2600-2606. DOI
23. Quan H, Zhang H, Zou L. Comparison of outcomes between patients with de-novo sinonasal squamous cell carcinoma vs malignant transformations from inverted papillomas. Int Forum Allergy Rhinol. 2020; 10:762-767. DOI
24. Li Y, Wang C, Wang R. Prognostic factors of sinonasal squamous cell carcinomas arising de novo and from inverted papilloma. Am J Rhinol Allergy. 2021; 35:114-121. DOI
25. Stepp WH, Farzal Z, Kimple AJ. HPV in the malignant transformation of sinonasal inverted papillomas: a meta-analysis. Int Forum Allergy Rhinol. 2021; 11:1461-1471. DOI
26. McCormick JP, Suh JD, Lee JT. Role of high-risk HPV detected by PCR in malignant sinonasal inverted papilloma: a meta-analysis. Laryngoscope. 2022; 132:926-932. DOI
27. Zhao RW, Guo ZQ, Zhang RX. Human papillomavirus infection and the malignant transformation of sinonasal inverted papilloma: a meta-analysis. J Clin Virol. 2016; 79:36-43. DOI
28. Sahnane N, Ottini G, Turri-Zanoni M. Comprehensive analysis of HPV infection, EGFR exon 20 mutations and LINE1 hypomethylation as risk factors for malignant transformation of sinonasal-inverted papilloma to squamous cell carcinoma. Int J Cancer. 2019; 144:1313-1320. DOI