The literature search was conducted on July 14, 2025. PubMed and Research Information Sharing Service (RISS) were searched, and all JR publications were screened for inclusion. Because JR was not indexed in PubMed from 1994 to 2021 [9], 786 articles were retrieved from RISS [10] and exported to EndNote [11]. Additionally, 102 articles published between 2022 and 2025 were retrieved from PubMed using the search strategy “Journal of Rhinology [Journal].” Two authors independently cross-checked the records and manually removed 52 duplicate articles available in both Korean and English. In total, 836 articles published between 1994 and 2025 were included in the analysis (Fig. 1).

Keywords were standardized prior to analysis to minimize redundancy and ensure conceptual consistency using a stepwise workflow. First, lexical normalization was performed to address superficial differences among keywords, including variations in plurality, spelling, and word form (e.g., “neoplasm” and “neoplasms” were unified under “neoplasm”). Similarly, identical concepts with minor linguistic variations were merged at this stage (e.g., “nasal septal perforation” and “nasal septum perforation” were unified under “nasal septal perforation”). Second, the normalized keywords were grouped according to the Medical Subject Headings (MeSH) database. Conceptually overlapping terms were consolidated under the official MeSH descriptor. This process enabled the unification of terms representing similar concepts, such as “SARS-CoV-2” and “Coronavirus.” The first two steps were conducted independently by two researchers and subsequently cross-checked to ensure consistency and validity. Third, to enhance clinical relevance, the keyword grouping decisions were independently validated by one author (JY Kim), an expert in otorhinolaryngology with more than 20 years of clinical experience and expertise in medical informatics. This reviewer reassessed the retrieved terms and their assigned groupings. For expert validation, a professor of otolaryngology reviewed the MeSH-consolidated keyword list and reclassified keywords with a frequency of five or more to ensure appropriate clinical contextualization. Keywords without an assigned group label were retained in their original form, whereas keywords assigned to a group were standardized using the representative group term. This process reduced the number of unique keywords from 120 original terms to 72 standardized groups (Supplementary Table 1 in the online-only Data Supplement). This expert-guided refinement was applied consistently in subsequent analyses to maintain clinical relevance while preserving less frequent, conceptually distinct terms.

Publications from 1994 onward (corresponding to Vol. 1 of JR) were used to analyze annual publication volume. Keyword analyses were restricted to publications from 2008 onward because RISS data completeness before 2008 could not be ensured. Korean keywords were translated into English to reduce redundancy, and duplicate keywords were removed.

The R bibliometrix package (version 5.1.0) [12] was used to analyze annual publication trends, keyword trends, and thematic evolution. The minimum keyword occurrence threshold was set at 5. Each keyword was represented by a dot indicating the median year of appearance and was sorted according to this value. Dot size reflected frequency, and lines connecting the first and third quartiles represented the primary period of discussion. For thematic evolution analysis, the top 250 keywords meeting the minimum frequency threshold were clustered using the “walktrap” algorithm, which identifies related keyword clusters based on random walks [13]. The period from 2008 to 2025 was divided into four intervals, and clusters predominating in each interval were identified. Thematic transitions across intervals were illustrated using directional arrows.

VOSviewer (version 1.6.20) [14] was used to conduct co-occurrence analysis of author-provided keywords. A bibliometric map was generated using the full counting approach, with all keywords as the unit of analysis. The minimum occurrence threshold was five. Two network visualizations were produced: 1) a network visualization illustrating keyword clusters and 2) an overlay visualization depicting temporal relationships based on average publication year. Each keyword was represented as a node, and each edge represented co-occurrence frequency. Link thickness reflected connection strength. The total link strength of a node was defined as the sum of its connection strengths. Clusters were distinguished by color and represented thematic domains. Denser areas indicated well-developed research domains, whereas sparser regions reflected emerging or niche topics. The average publication year was used to identify emerging themes.

The number of links, total link strength, and number of occurrences were obtained using VOSviewer. Closeness centrality and betweenness centrality values for each keyword were calculated through the “network overview” tab and then exported using the “data laboratory” feature in Gephi (version 0.10.1) [15] to assess their network properties. Closeness centrality, betweenness centrality, and the average publication year data were calculated and extracted using Gephi’s native “network diameter” plug-in provided in the “network overview” tab. Together, these metrics provided a comprehensive characterization of the research network (Table 1).

We analyzed JR’s annual scientific output from 1994 to 2025 (Fig. 2). The annual publication volume averaged 26.6 articles per year through 2024. The highest annual output occurred in 2009 (35 articles), whereas the lowest was recorded in 2001 (13 articles). Over the most recent five-year period, the average increased to 29.4 articles per year, indicating an upward trend in recent publication activity.

The frequency of each keyword’s appearance reflects the main research interests of JR. We examined the most frequently used keywords in JR from 1994 to 2025 and identified the top 15 keywords (Fig. 3). The most frequently appearing keyword was “paranasal sinus” (n=85), followed by “sleep apnea syndromes” (n=74), “sinusitis” (n=71), and “endoscopy” (n=63).

By identifying the periods during which keywords were most actively used, we can infer JR’s research priorities and how these priorities changed over time. We analyzed the first and third quartiles of each keyword’s occurrence period to represent the interval during which the keyword was predominantly cited in JR (Fig. 4). Based on these results, we examined high-frequency terms longitudinally. Around 2015, “sinusitis” and “endoscopy” showed high frequencies and were consistently observed for approximately 10 years (approximately 2011–2021). In addition, “paranasal sinus” appeared frequently around 2016 and was sustained for a relatively long period (2010–2020). Around 2018, “sleep apnea syndromes” became prominent and continued to appear from 2012 to 2022. In contrast, keywords such as “Caldwell–Luc operation” and “epithelial cells,” which were more common in the early 2010s, were not sustained in later years. More recently, after 2020, methodological keywords such as “meta-analysis” and “surveys and questionnaires” newly emerged.

From 2008 to 2025, keywords were clustered using the walktrap algorithm, and the most frequently discussed clusters in each period were identified (Fig. 5). “Rhinitis” and “sleep apnea syndromes” appeared across all periods, indicating that these clusters have been consistently addressed in JR. From 2008 to 2012, keywords related to nasal anatomical structures and diseases directly associated with the nose (e.g., “nasal cavity” and “paranasal sinus”) predominated. From 2013 to 2016, this pattern remained similar, although diagnostic technique–related keywords such as “computed tomography” were added. From 2017 to 2021, keywords related to infectious diseases (e.g., “SARS-CoV-2”) and procedural topics (e.g., “rhinoplasty”) emerged. From 2022 to 2025, surgical method–related keywords, including “endoscopy” and “septoplasty,” became prominent. Overall, basic disease-related keywords such as “nasal polyps” and “neoplasms” were dominant in earlier periods, whereas surgical and procedural keywords such as “rhinoplasty” and “septoplasty” became more prominent in recent years.

A total of 1,019 keywords were identified, of which 49 met the threshold of 5 minimum occurrences (Fig. 6). The network demonstrated seven clusters with 208 links and a total link strength of 395. The first cluster (9 terms), colored red, focused on nasal cavity anatomy and obstruction, including epistaxis, turbinate surgery, and SARS-CoV-2. The second cluster (7 terms), colored green, concentrated on inflammatory and immune responses in rhinitis, asthma, and nasal polyps. The third cluster (7 terms), shown in dark blue, encompassed skull base pathology and reconstruction, including neoplasms and cerebrospinal fluid. The fourth cluster (7 terms), colored yellow, focused on pediatric conditions and sleep apnea syndromes alongside surgical interventions such as tonsillectomy and rhinoplasty. The fifth cluster (6 terms), colored purple, addressed paranasal sinus pathology, particularly sinusitis and fungal infections such as aspergillosis. The sixth cluster (5 terms), visualized in light blue, encompassed endoscopic nasal surgery and associated symptoms, including facial pain and headache. The seventh cluster (5 terms), colored orange, focused on nasal septal structural conditions and septoplasty. The final cluster (3 terms), colored brown, concentrated on surgical treatment outcomes and the use of computed tomography.

Among the analyzed keywords, “paranasal sinus” appeared most frequently (n=85), followed by “sleep apnea syndromes” (n=74), “sinusitis” (n=71), and “endoscopy” (n=63). In betweenness centrality, “endoscopy” ranked highest (262.342), and “sinusitis” (191.119) and “paranasal sinus” (143.650) also showed strong bridging roles within the co-occurrence network. For closeness centrality, “endoscopy” ranked highest (0.750), followed by “sinusitis” (0.706), “paranasal sinus” (0.696), and “nasal septum” (0.574), indicating that these terms served as central connectors across research themes. The terms with the highest number of links were “endoscopy” (n=32), “sinusitis” (n=28), “paranasal sinus” (n=27), and “nasal cavity” (n=18), indicating broad relevance across the network. The terms with the highest total link strength were “paranasal sinus” (n=92), “endoscopy” (n=82), “sinusitis” (n=82), “nasal cavity” (n=35), and “nasal septum” (n=28). High total link strength indicates strong connectivity with multiple research clusters, suggesting that these terms were well integrated across thematic domains.