In this study, all literature data were obtained from the Web of Science Core Collection database. This database encompasses academic publications from over 250 diverse global disciplinary fields and stands as a widely utilized platform for bibliometric analysis among research scholars (14–16). The bibliometric analysis in this article covers the period from January 1, 2004, to April 9, 2024. The search query employed was as follows: TS= (“papillary thyroid carcinoma” OR “thyroid papillary carcinoma” OR “thyroid carcinoma, papillary” OR “papillary thyroid cancer” OR “thyroid papillary cancer” OR “thyroid cancer, papillary”) AND TS= (Prognosis OR Prognoses OR Prognostic). Only English-language publications from the search results were included, and after excluding literature types such as letters, comments, and conference papers, a total of 3430 relevant articles were obtained. The detailed literature screening process is shown in Figure 1.

The study utilized various analysis tools, including Microsoft Office Excel 2021, VOSviewer (version 1.6.18), CiteSpace (version 6.1.R6), and the R package “bibliometrix,” for data analysis and visualization. Prior to analysis, we manually performed data cleaning and disambiguation to merge synonymous nodes and standardize author names and institutional affiliations (e.g., merging different spellings of the same institution, such as “Shanghai Jiao Tong University” and “Shanghai Jiao Tong Univ”) to ensure the accuracy of frequency counts and network linkages. Initially, data related to 1108 articles on papillary thyroid cancer (PTC) prognosis were uniformly extracted from the Web of Science Core Collection database, comprising information such as titles, authors, keywords, affiliations, countries/regions, citations, journals, and publication dates. This dataset underwent preliminary screening and processing before being analyzed via VOSviewer, CiteSpace, and the R package “bibliometrix” for bibliometric analysis. CiteSpace, known for creating knowledge maps in specific fields, was utilized for co-occurrence and clustering analysis of authors, research institutions, and countries (17, 18). VOSviewer, a bibliometric analysis software, facilitates the extraction and analysis of key information from publications, including collaboration relationships among countries, authors, institutions, and co-occurrence patterns of keywords (19, 20). For author and institution citation counts presented in Tables 1, 2, self-citations were not excluded, as retaining them provides a more comprehensive reflection of overall academic influence within the field, consistent with common practice in bibliometric analysis. Finally, Bibliometrix, an open-source R package, was employed for comprehensive bibliometric and scientometric analysis, particularly for analyzing evolving trends of keywords in the literature (21).

Figure 2A illustrates the upward trends in both publication and citation counts from 2004 to 2024. Both metrics consistently increased during this period. Significant increases in publications were observed in 2010, 2012, and 2020, with a peak of 325 publications per year in 2022. Similarly, citation counts experienced notable growth during 2011–2014 and 2017-2020, likely attributed to influential publications preceding these periods. The highest citation peak, reaching 10,219 citations, was also recorded in 2022.

Furthermore, a polynomial fit was applied to the cumulative annual publication count, as depicted in Figures 2B, with the fitting equation y=0.0774x³+7.2983x²-10.613x+68.36 and a fitting goodness of R² = 0.9996.

Analysis of countries/regions contributing to the PTC prognosis literature offers insights into the global distribution of research outcomes and focus areas (Table 3). China and the United States have emerged as key players in this field, leading in both article production and citation rates. While China has a significantly higher number of articles (1279) than other countries, the United States surpasses in total citation frequency (26,954), indicating potentially greater influence. Notable contributions are also observed from countries such as South Korea, Italy, and Japan, further enriching the research landscape in PTC prognosis.

Using VOSviewer, we performed an extensive analysis of countries/regions. The collaboration dynamics among these nations are depicted in Figure 3 via a chord diagram, with each country represented by differently colored ribbons reflecting collaboration strength. The findings highlight China as a leader in global collaboration, closely followed by the United States. Strong collaboration is observed between these two nations and between the United States and Italy. South Korea also plays a significant role in fostering academic exchanges, particularly with the United States.

Table 1 presents the top ten authors on the basis of publication count and total co-citations. Leading contributors such as Miyauchi, Akira, Ito, Yasuhiro, Miya, and Akihiro have each authored over 50 publications in the field, notably all affiliated with Kuma Hospital in Japan. Regional diversity is evident among the top authors by publication count, with representations from South Korea, Japan, the USA, and China, which is consistent with the analysis of research dominance among countries/regions. Notably, the majority of authors with the highest total co-citations are from the USA, with only one exception from South Korea. Authors from institutions such as Johns Hopkins University and the University of Pittsburgh stand out among the top contributors.

To further illustrate the connections among these authors and their potential research similarities, the results of the analysis using VOSviewer are presented in Figure 4. In Figures 4A, different colors distinguish several major clusters, visualizing the collaboration relationships between authors within the same cluster. In Figures 4A, the largest author collaboration cluster, represented in green, centers around authors from Kuma Hospital in Japan, such as Miyauchi, Akira and Miya, Akihiro. Another significant cluster, shown in purple on the left, mainly includes authors from the University of Ulsan in South Korea, such as Chung, Ki-Wook and Kim, Won Bae. Various clusters of South Korean authors are visible, indicating geographical proximity and interconnectivity among them. Another notable cluster, highlighted in orange, primarily consists of Chinese authors, whereas the yellow cluster represents collaboration among American authors. Geographical proximity plays a role in author collaboration, but international collaboration is also significant, as seen in the red cluster with multinational collaboration, including authors from Italy, the USA, and Australia. Overall, the distribution of author collaboration clusters reveals spatial patterns, with most collaborations occurring among authors from the same country or institution, resulting in relatively independent clusters with minimal connections between them.

Figure 4B provides a visual representation of co-citation relationships between authors, indicating how frequently two authors are cited together by third-party authors in the same publication. This indirect co-citation relationship offers insights into potential connections and research relevance between these authors. While authors in this field cover a broad range of research topics, including neuroscience and the respiratory system, there are also more focused research directions within this broader scope. The authors are mainly divided into four clusters distinguished by color. In the red cluster on the right, the emphasis is primarily on clinical and surgical aspects. Authors such as Ito, Yasuhiro, Haugen, Br, Mazzaferri, El, and Davies, L have made significant contributions in this direction. The green cluster on the left predominantly focuses on cellular and molecular biology, featuring authors such as Xing, Mingzhao, Nikiforov, Ye, and Elisei, Rossella. Authors in the blue cluster in the middle primarily engage in pathology, laboratory medicine, or experimental medicine, including Baloch, Zubair W., Rosai, Juan, and Livolsi, V. A. The yellow cluster at the bottom encompasses authors with more diverse research directions, yet their studies also exhibit common themes in endocrinology, metabolism, and oncology.

Table 2 presents the top ten institutions in terms of the number of publications, total link strength, and total citations. Kuma Hospital in Japan has emerged as the most prolific institution in output in the field of PTC prognosis, with 85 publications, and it also ranks among the top three in terms of total citations (4645). The second most common institution is Yonsei University in South Korea, with 84 publications, while three institutions from China—Shanghai Jiao Tong University (78 publications), China Medical University (73 publications), and Zhejiang University (72 publications)—also make significant contributions to PTC prognosis research. Notably, Johns Hopkins University in the USA stands out as the institution with the highest number of total citations (5508), followed by Memorial Sloan Kettering Cancer Center (4912).

Figure 5 provides a visual representation of collaborative relationships among institutions, illustrating the spatial distribution of these collaborations. In Figures 5A, the upper left green cluster primarily gathers South Korean institutions with close collaborative ties, including several institutions that have made significant contributions to the field, such as Sungkyunkwan University, Seoul National University, and Yonsei University. The pink and red clusters on the right exhibit greater diversity in regions, predominantly featuring institutions from Europe and the USA, including Karolinska University Hospital in Sweden and the University of Porto in Portugal, as well as renowned American universities or medical centers such as Harvard Medical School, Memorial Sloan Kettering Cancer Center, and University of Texas MD Anderson Cancer Center. Although the brown cluster comprises fewer institutions, the University of Pisa, which has the strongest total link strength, is included in this cluster. Additionally, other Italian institutions such as the University of Naples Federico II and University of Padua are also present. The purple, yellow, and dark blue clusters at the bottom represent institutions in China, including influential institutions such as Shanghai Jiao Tong University, Zhejiang University, Huazhong University of Science and Technology, and Sun Yat-sen University. While the light blue and orange clusters also include Chinese institutions, they mainly depict multinational collaborations, such as academic exchanges between Kuma Hospital in Japan and Shandong University, and between the University of Sydney and the University of Milan.

Figures 5B, which extends the spatial analysis in Figures 5A, introduces a temporal dimension by examining the frequency of collaboration over the past five years. The color scheme highlights the chronological sequence of collaboration relationships on the basis of their occurrence frequency. This reveals that in the early years of this century, research on PTC prognosis was primarily conducted by institutions from Japan, South Korea, Europe, and the USA, with establishments such as Kuma Hospital and the University of Texas MD Anderson Cancer Center laying a solid foundation for the future development of this field. However, in recent years, there has been a noticeable shift, with academic collaboration among institutions in China injecting new vitality into this field.

Table 4 presents the top ten journals in terms of the number of publications and citations in this field. The journal with the highest number of publications is Thyroid (173 articles), which also has the highest citation count (9121 citations). Following Thyroid, we have Frontiers in Endocrinology (92 articles), World Journal of Surgery (85 articles), and Cancers (83 articles). The journals with the second highest citation count are Journal of Clinical Endocrinology and Metabolism (7532 citations) and World Journal of Surgery (3815 citations), indicating their central role in this field. More than half of the journals listed in the top rankings for both publication and citation frequency are high-impact journals in the Q1 or Q2 category. Given that the impact factor is an important indicator of journal quality, it can be inferred that articles on PTC prognosis published in these journals maintain a high level of quality.

The citation and collaboration relationships among these journals are visually depicted in Figure 6A, offering a more intuitive understanding of their interactions. Thyroid prominently emerges as a core journal in the field and is closely linked to World Journal of Surgery, Surgery, and Endocrine Journal. Figure 6B visualizes the research domains represented by these journals through distinct colored clusters. For example, journals focusing on medicine and surgery, such as the Journal of Clinical Endocrinology and Metabolism, Surgery, and World Journal of Surgery, are grouped in the green cluster. Journals covering research content in the fields of molecular biology, cell biology, and pathology, such as Acta Cytologica, Cancer Cytopathology, Endocrine Pathology, and The Journal of Pathology, are clustered in the blue category. Journals related to cancer research or endocrinology, including Endocrine-Related Cancer, Oncotarget, and CA: A Cancer Journal for Clinicians, are highlighted in the red or yellow clusters.

Figure 6C provides additional clarification on the citation and co-citation relationships among journals through an overlay analysis obtained using CiteSpace. The citing journals are listed on the right, whereas the cited journals are listed on the left. The thickness of the lines between them indicates the strength of the citation relationship. It is apparent that the citing journals primarily originate from fields such as medicine, molecular biology, immunology, and clinical studies, whereas the cited journals predominantly belong to domains such as molecular genetics, health, and medicine.

Table 5 presents the top 20 keywords based on occurrence frequency and total link strength, which represent the main directions or core viewpoints of the research and can accurately reflect the research hotspots and frontiers in the field during a specific time period. The most frequently occurring keyword is papillary thyroid carcinoma (1845 occurrences), followed by thyroid cancer (524 occurrences) and prognosis (444 occurrences). Additionally, there are keywords related to the pathological manifestations and treatment targets of PTC, such as lymph node metastasis, biomarker, and BRAF V600E.

Figures 7A, B depict the co-occurrence relationships and strengths among key terms, not only reflecting the overlap of research content but also aiding in the identification of potential connections between research topics and paradigms. This is valuable for researchers to explore new research directions. As shown in Figure 7A, the key term with the highest co-occurrence strength is papillary thyroid carcinoma, which is located at the center. It forms a red cluster with some terms related to cell biology and molecular biology on the right, such as microRNA, proliferation, apoptosis, and biomarker. The yellow cluster below mainly consists of terms related to cancer, such as TCGA (The Cancer Genome Atlas), anaplastic thyroid carcinoma, follicular thyroid cancer, etc. The orange cluster mainly lists terms related to risk factors and disease progression of PTC, such as risk factors, lateral node metastasis, risk stratification, etc. The keywords in the deep blue and light blue clusters located at the bottom left are mainly related to clinical medicine and surgery, such as thyroidectomy, radioactive iodine, surgery. The keywords in the green cluster focus on describing the pathological manifestations of PTC and clinical or laboratory diagnostic methods, such as pathology, ultrasonography, fine-needle aspiration, etc. The purple cluster displays some terms related to cell biology or molecular biology related to PTC, such as immunohistochemistry, BRAF V600E, etc. Figure 7B includes a description of the contribution strength of keywords. It is evident that in recent years, the high-temperature keywords mainly describe nouns related to oncology, such as anaplastic thyroid carcinoma, follicular thyroid cancer, medullary thyroid carcinoma, etc.

Figure 7C presents the top 25 keywords exhibiting the most robust citation bursts. The keywords with the highest burst strengths are prognostic factors (34.18) and high prevalence (23.06), indicating their early emergence as research hotspots. The keyword with the longest duration of sustained research interest is differentiated carcinoma (2004-2015). Below, there are also several keywords still experiencing bursts, demonstrating that research on these topics remains current hotspots, such as guidelines (13.14), association guidelines (12.84), PTC (papillary thyroid carcinoma) (12.26), and case report (8.91).

The citation count is an important indicator for assessing the quality and impact of articles. Highly cited articles play a crucial role in driving research progress and reflecting research trends. Table 6 presents basic information on the top fifteen highly cited articles, with the highest citation count attributed to the review by Prahallad, A et al., published in 2012 in Nature, titled “Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR” (22) (1499 citations). This article primarily addresses the issue of the poor efficacy of the drug PL4032 in patients with BRAF (V600E) mutant colorectal cancer. It analyzes the causes and mechanisms of action of BRAF (V600E), proposing a combination therapy using BRAF and EGFR inhibitors was proposed to treat patients with BRAF (V600E) mutant colorectal cancer. The second most highly cited literature, with 764 citations, is the article titled “BRAF mutation predicts a poorer clinical prognosis for papillary thyroid cancer” (23) authored by Xing, MZ et al. and published in the Journal of Clinical Endocrinology and Metabolism in 2005. Given the early publication of this article, we can infer that its research findings laid the groundwork for many subsequent studies. Through a multicenter study, this study explored the associations between BRAF mutations and the occurrence of extrathyroidal invasion, lymph node metastasis, and advanced tumor stages III/IV in patients with PTC. These findings underscore the importance of BRAF in disease diagnosis, progression, and its potential role as a therapeutic target. In the third article, also authored by Xing, MZ et al. and published in 2013 in JAMA - Journal of the American Medical Association, titled “Association Between BRAF V600E Mutation and Mortality in Patients With Papillary Thyroid Cancer” (24), the author further investigated the impact of BRAF on patients with PTC. Through a retrospective multicenter study, they reported an association between the BRAF (V600E) mutation and increased mortality in patients with PTC, highlighting the significant regulatory role of BRAF in the prognosis of PTC patients.

Figure 8A illustrates the co-citation patterns of the literature related to PTC prognosis. By connecting clusters, it delineates the developmental and evolving sequence of core concepts within these highly cited publications. The largest cluster, labeled as #0 nomogram, is noteworthy as it is the latest-formed cluster in the graph and its connections extend to multiple clusters, indicating that research on PTC prognosis from this perspective has emerged and evolved as an important research direction since the early 21st century, integrating and developing various research contents over time. Key concepts influencing this cluster include #4 TNM (TNM Classification of Malignant Tumors), #3 NIFTP (Noninvasive follicular thyroid neoplasm with papillary), #2 BRAF, and #1 thyroidectomy. In terms of the overall developmental trend of research content, the early concept #15 staging gradually materializes into the more mature #4 TNM. Research on PTC has evolved from # childhood PTC and early surgical treatment methods # thyroidectomy to delving deeper into the disease mechanisms, mainly based on #2 BRAF as a therapeutic target for developing novel intervention strategies.

Figure 8B illustrates the reference with the highest burst intensity, which is the work by Haugen BR, published in 2016. Titled “2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer” (25), it has a burst intensity of 163.16. One year after publication, this reference gained widespread attention and continues to attract significant interest (2017–2021). Cooper DS’s work, published in 2009 and titled “Revised American Thyroid Association Management Guidelines for Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association (ATA) Guidelines Taskforce on Thyroid Nodules and Differentiated Thyroid Cancer” (8), follows closely with a burst intensity of 95.42. Furthermore, the relatively recent references “Thyroid cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up” (9) by Filetti S et al. and “Overview of the 2022 WHO Classification of Thyroid Neoplasms” (26) by Baloch ZW et al., are still experiencing bursts, indicating the ongoing interest and impact of these articles.