Hotspots and trends of risk factors in gastric cancer: A visualization and bibliometric analysis

Abstract

BACKGROUND

The lack of specific symptoms of gastric cancer (GC) causes great challenges in its early diagnosis. Thus it is essential to identify the risk factors for early diagnosis and treatment of GC and to improve the survival rates.

AIM

To assist physicians in identifying changes in the output of publications and research hotspots related to risk factors for GC, constructing a list of key risk factors, and providing a reference for early identification of patients at high risk for GC.

METHODS

Research articles on risk factors for GC were searched in the Web of Science core collection, and relevant information was extracted after screening. The literature was analyzed using Microsoft Excel 2019, CiteSpace V, and VOSviewer 1.6.18.

RESULTS

A total of 2514 papers from 72 countries and 2507 research institutions were retrieved. China (n = 1061), National Cancer Center (n = 138), and Shoichiro Tsugane (n = 36) were the most productive country, institution, or author, respectively. The research hotspots in the study of risk factors for GC are summarized in four areas, namely: Helicobacter pylori (H. pylori) infection, single nucleotide polymorphism, bio-diagnostic markers, and GC risk prediction models.

CONCLUSION

In this study, we found that H. pylori infection is the most significant risk factor for GC; single-nucleotide polymorphism (SNP) is the most dominant genetic factor for GC; bio-diagnostic markers are the most promising diagnostic modality for GC. GC risk prediction models are the latest current research hotspot. We conclude that the most important risk factors for the development of GC are H. pylori infection, SNP, smoking, diet, and alcohol.

Key Words: Gastric cancer, Risk factor, Bibliometric, Research hotspots, VOSviewer

Core Tip: In this study, we searched and organized relevant articles in the field of risk factors for gastric cancer (GC) and carried out a bibliometric analysis. In this manuscript, we summarize the current state of development in the field, produce a table of key risk factors in GC, and achieve a prediction of future research trends.

INTRODUCTION

Gastric cancer (GC) is a prevalent cancer worldwide, with over 1 million new cases and 783000 new deaths worldwide in the year 2018 alone[1]. According to the World Health Organization (WHO), GC is the fifth most prevalent cancer and the fourth leading cause of cancer-related deaths worldwide as of 2020[2]. In China, GC ranks as the third leading cause of cancer-related deaths[3]. Although surgery is the standard treatment for GC, early detection, diagnosis, and treatment are pivotal in improving patient outcomes[4]. The early symptoms of GC patients are not obvious, leading to diagnosis at advanced stages[5]. It has been noted that patients diagnosed with early-stage GC can achieve a 5-year survival rate of up to 90%[6], while for those diagnosed with advanced stage GC, the rate is only 10%[7]. In East Asia, population-based screening programs have been implemented in countries with high GC morbidity and mortality. For example, opportunistic screening has been conducted in China, Japan, Korea, and Singapore[8], and in particular, Japan and Korea have implemented national GC screening programs[9]. The Chinese Ministry of Health launched a program to promote early diagnosis and treatment of cancer in rural areas with a high prevalence of malignancies in 2005, of which GC is a key target[10]. Evidence from a nested case-control study in Korea demonstrated that endoscopic screening program can reduce GC-related mortality by 47%[11]. These screening programs increased the detection rate of early GC and significantly reduced the mortality rate of GC patients[4]. However, despite advancements in the assessment and management of gastric precancerous lesions, significant uncertainty persists regarding risk factors indicating tumor progression[12]. Therefore, the identification of risk factors is of great importance for GC screening. For example, compared to non-smoking men, those with a smoking history had 1.13 and 2.43 times higher risks of GC when presenting with normal and below-normal serum pepsinogen I (PGI) levels, respectively[13]. Adamu et al[14] found that advanced age was a key risk factor for the development and progression of chronic atrophic gastritis. In addition, factors such as alcohol consumption[15] and a family history of tumors[16] also increase the risk of GC development.

In order to better understand the field, the following questions need to be addressed first: (1) What are the annual trends in articles? (2) What are the differences in nationals, institutionals and authors contributions? (3) What are the most popular risk factors in research? And (4) In what area is future research likely to emerge? Traditional literature reviews are time-consuming and labor-intensive, and there are limitations to the research[17], making it difficult to provide a reasonable description of the aforementioned issues. In contrast, bibliometrics, through a multidisciplinary approach[18], can achieve qualitative and quantitative analyses of publications in a specific field[19] so as to sort out the research lineages and grasp the research trends and directions[20].

There is no research to date that has sorted and analyzed risk factors for GC-related research results from a bibliometric perspective. This paper provides a detailed analysis of risk factors for GC field in terms of countries, institutions, authors, sources and keywords, allowing physicians to stay up-to-date with the latest research evidence in key clinical areas and make increasingly informed decisions.

MATERIALS AND METHODS

Search strategies

Data were retrieved using the Science Citation Index-Expanded in the Web of Science Core Collection (WOSCC). In order to keep the proportion of articles that were not relevant to the topic within a reasonable range, the search method of title search was utilized[21-23]. The search formula was set as: TI = (“Stomach Neoplasm” OR “Gastric Neoplasm” OR “Cancer of Stomach” OR “Stomach Cancer” OR “Gastric Cancer” OR “Cancer of the Stomach”) AND TI = (“Risk Factor” OR “Social Risk Factor” OR “Health Correlates” OR “Population at Risk” OR “Risk Score” OR “Risk Factor Score” OR “Predictor” OR “Risk”).

Inclusion criteria

The screening criteria were: (1) Publication time range: from Jan. 01, 2000 to Dec. 04, 2022; (2) publication language in English; and (3) publication type was research article or review. The data retrieval strategy and screening process of this study are shown in Figure 1.

Figure 1 Flow chart of literature screening.

Data collection and analysis

In this study, Microsoft Excel 2019, CiteSpace V, and VOSviewer 1.6.18 software were used for data analysis. Microsoft Excel software allows descriptive statistics on the number of publications and citations by countries, institutions, and authors[24], where Impact factor (IF), and Journal Citation Report (JCR) category[25] were obtained using JCR Science Edition (2021). VOSviewer enables the visual analysis of collaborative networks of countries, institutions and authors as well as keyword clusters. CiteSpace allows for keyword burst detection, which leads to the analysis of research hotspots and trends in the time dimension[26]. The software parameters were set as time slice (1990-2022), years per slice (1), and selection strategy (g-index, k = 25)[27].

Quality control

Two investigators (Li M and Gao N) performed the data search independently and discussed any discrepancies. Data retrieval and export were completed on the same day (December 04, 2022), minimizing bias caused by database updates. This study did not include any animals or experiments and therefore did not require ethics committee approval.

RESULTS

Publication outputs and citation trend

Between 2000 and 2022, WOSCC published 2514 papers on risk factors for GC, including 2352 articles and 162 reviews. Figure 2 illustrates the annual growth trend of publication volume and citations, which showed an overall upward trend in the past 20 years and reached its peak in 2021 (n = 217). The exception to this trend was a small decrease in volume in 2018 and 2019. The annual growth curve of publications (y = 26.911e^0.0996x) was fitted by an exponential function, where R² = 0.9033, suggesting a strong correlation between the number of publications and the year. For the average annual citations of publications, there were two peaks in 2000 and 2003, respectively, after which the average citations decreased year by year. It can be predicted that the field will continue to attract the attention of scholars in the coming years, meanwhile the number of annual publications will continue to grow.

Figure 2 Annual number of the publications in risk factors for gastric cancer research.

Contributions of countries

Seventy-two countries were involved in publishing risk factors for GC-related studies, and China was the main contributing country in this field with 1061 publications, accounting for 42.20% of the total publications, followed by Japan (n = 467), United States (n = 383), and South Korea (n = 311). Nine of the top ten contributing countries were developed countries, indicating a strong correlation between a country’s economic development level and its publication output. In terms of citations, the United Kingdom has the highest average number of citations (n = 73.60), followed by Netherlands (n = 57.31), and United States (n = 55.34), respectively. This may be a reflection of the differences in research quality between countries. Although China is a high-output country in risk factors for GC, the quality of research needs to be further improved (Table 1). Worldwide, the participation of East Asian, North American, and Western European countries is high, while the participation of African, Western Asian, and Eastern European countries is low, with some regional differences in participation (Figure 3A). Among the cooperative networks, the United States has the highest total link strength (TLS = 644) and has cooperated with 45 countries, among which the closest cooperation is with China (TLS = 131). Meanwhile, it can be seen by the node color that European countries such as France, Switzerland, and Norway conducted related research around 2011, while China and Korea started later, but developed faster (Figure 3B).

Figure 3 Depiction of countries involved in risk factors for gastric cancer research. A: Distribution of risk factors for gastric cancer publications in the world map; B: Overlay visualization map of co-authorship among countries.

Table 1 The top 10 most productive countries in research.

Rank	Country	TP	Percent (%)	TC	CPP
1	China	1061	42.20	22402	21.11
2	Japan	467	18.58	13585	29.09
3	United States	383	15.23	21194	55.34
4	South Korea	311	12.37	6831	21.96
5	Italy	152	6.05	7327	48.20
6	Sweden	122	4.85	5669	46.47
7	Germany	116	4.61	4919	42.41
8	United Kingdom	108	4.30	7949	73.60
9	Spain	89	3.54	4225	47.47
10	Netherlands	83	3.30	4757	57.31

Total publications: Total number of publications during the observation period; Total citations: Total citations of publications; CPP: Number of citations per publication; TP: Total publications; TC: Total citations; CPP: TC/TP.

Contributions of institutions

There were 2507 institutions involved in the publication of risk factors for GC-related articles. The highest contributor was the National Cancer Center with 138 publications, followed by Nanjing Medical University (n = 122), National Cancer Institute (n = 104), and Seoul National University (n = 88). 7 of the top 10 contributors are located in Asia, including 4 in China and 3 in Korea. In addition, National Cancer Institute ranks first in terms of total citations and average citations (Table 2). The National Cancer Institute has the highest TLS (TLS = 236) in the collaborative network, having collaborated with 152 institutions, with the closest collaboration with Vanderbilt University (TLS = 25) (Figure 4). Both institutions have carried out extensive research work on the correlation between diet and GC and found a correlation between GC and the folic acid content of food[28], meat intake[29], and fruit intake[30].

Figure 4 Overlay visualization map of co-authorship among institutions with 10 or more publications.

Table 2 The top 10 most productive institutions in gastroparesis research.

Rank	Institution	Country	TP	TC	CPP
1	National Cancer Center	Korea	138	5475	39.67
2	Nanjing Medical University	China	122	2906	23.82
3	National Cancer Institute	United States	104	9330	89.71
4	Seoul National University	Korea	88	2336	26.55
5	China Medical University	China	81	1460	18.02
6	Karolinska Institute	Sweden	73	3111	42.62
7	Fudan University	China	57	1435	25.18
8	Shanghai Jiao Tong University	China	57	1706	29.93
9	Yonsei University	Korea	53	1628	30.72
10	Vanderbilt University	United States	50	2843	56.86

Total publications: Total number of publications during the observation period; Total citations: Total citations of publications; CPP: Number of citations per publication; TP: Total publications; TC: Total citations; CPP: TC/TP.

Journal analysis

There are 489 journals with publications in the field of risk factors for GC, and Table 3 shows the 10 journals with the highest number of articles. These 10 journals published 604 articles, accounting for 24.03% of the total number of articles. International Journal of Cancer published the most articles with 115, followed by Gastric Cancer, Plos One, and World Journal of Gastroenterology, all with 71 articles. Only two journals had IFs >5, Gastric Cancer (IF = 7.70) and World Journal of Gastroenterology (IF = 5.37). Three journals were temporarily outside the SCI category and two journals were in the Q1 JCR division.

Table 3 The top 10 Journals with the largest number of publications in research.

Rank	Journal	TP	TC	CPP	JCR	IF2021
1	International Journal of Cancer	115	6620	57.57	Q2	4.37
2	Gastric Cancer	71	1751	24.66	Q1	7.70
3	Plos One	71	1417	19.96	Q2	3.75
4	World Journal of Gastroenterology	71	2364	33.30	Q2	5.37
5	Asian Pacific Journal of Cancer Prevention	56	742	13.25	-	-
6	Cancer Epidemiology Biomarkers & Prevention	55	3968	72.15	-	-
7	Annals of Surgical Oncology	42	1350	32.14	Q1	4.34
8	BMC Cancer	42	843	20.07	Q2	4.64
9	Oncotarget	41	849	20.71	-	-
10	Medicine	40	605	15.13	Q3	1.82

Total publications: Total number of publications during the observation period; Total citations: Total citations of publications; CPP: Number of citations per publication; TP: Total publications; TC: Total citations; CPP: TC/TP.

Authors and co-cited authors

A total of 12569 researchers were involved in publishing risk factors for GC-related articles, with Il Ju Choi publishing the most articles (n = 36), followed by Shoichiro Tsugane (n = 34), Jeongseon Kim (n = 31), and Yuan Yuan (n = 31). Wong-ho Chow has the highest CPP value along with a relatively high study quality (Table 4). A threshold value of 10 was set to construct an author collaboration network, in which Shoichiro Tsugane had the highest TLS (TLS = 169) and has successively collaborated with 70 authors. In addition, it can be seen by the node color that numerous tightly knit research teams are involved in the risk factors for GC field (Figure 5A). A total of 30894 researchers constitute the co-cited authors of publications in the risk factors for GC field, with P Correa ranked first in terms of co-citations (n = 617), followed by Dm Parkin (n = 442) and Ca Gonzalez (n = 393). A threshold value of 50 was set to construct the co-cited authors' network, and the results are presented in Figure 5B.

Figure 5 The distribution of authors and co-cited authors in risk factors for gastric cancer research. A: Network visualization map of authors with 10 or more publications; B: Visualization map of co-cited authors.

Table 4 The top 10 authors and co-cited authors in risk factors for gastric cancer research.

Rank	Author	TP	TC	CPP	Co-cited author	Co-citations
1	Il Ju Choi	36	706	19.61	P Correa	617
2	Shoichiro Tsugane	34	1021	30.03	Dm Parkin	422
3	Jeongseon Kim	31	619	19.97	Ca Gonzalez	393
4	Yuan yuan	31	526	16.97	J Ferlay	352
5	Christian C Abnet	30	1161	38.70	Em El-omar	271
6	Wong-ho Chow	23	1091	47.43	F Bray	270
7	Ping Li	22	336	15.27	A Jemal	241
8	Xiao-ou Shu	22	564	25.64	N Uemura	233
9	Li Yang	22	414	18.82	M Rugge	230
10	Wei Zheng	22	564	25.64	P Lauren	228

Total publications: Total number of publications during the observation period; Total citations: Total citations of publications; CPP: Number of citations per publication; TP: Total publications; TC: Total citations; CPP: TC/TP.

Top cited publications and co-cited references

Highly cited papers indicate a significant impact in a field[31], reflecting the hot spots and depth of research in the field[32]. Table 5 lists the top 10 cited publications, with Emad M El-Omar's "Interleukin-1 polymorphisms associated with increased risk of gastric cancer" published in 2000 being the most cited. Two of these ten articles are reviews on risk factors for GC and three are related to Helicobacter pylori (H. pylori) infection. Further co-cited reference analysis was performed, and the results are shown in Table 6.

Table 5 The top 10 documents in citation analysis of publications in risk factors for gastric cancer research.

Rank	Title	First author	Source	Publication year	TC
1	Interleukin-1 polymorphisms associated with increased risk of gastric cancer	Emad M El-Omar	Nature	2000	1800
2	Helicobacter pylori eradication to prevent gastric cancer in a high-risk region of China: a randomized controlled trial	Benjamin Chun-Yu Wong	Jama	2004	1045
3	Gastric cancer: descriptive epidemiology, risk factors, screening, and prevention	Parisa Karimi	Cancer Epidemiol Biomarkers Prev	2014	1015
4	Helicobacter pylori and gastric cancer: factors that modulate disease risk	Lydia E Wroblewski	Clin Microbiol Rev	2010	811
5	Increased risk of noncardia gastric cancer associated with proinflammatory cytokine gene polymorphisms	Emad M El-Omar	Gastroenterology	2003	711
6	Population attributable risks of esophageal and gastric cancers	Lawrence S Engel	J Natl Cancer Inst	2003	513
7	Prospective study of risk factors for esophageal and gastric cancers in the Linxian general population trial cohort in China	Gina D Tran	Int J Cancer	2005	494
8	Gastric cancer risk in patients with premalignant gastric lesions: a nationwide cohort study in the Netherlands	Annemarie C de Vries	Gastroenterology	2008	468
9	Progression of chronic atrophic gastritis associated with Helicobacter pylori infection increases risk of gastric cancer	Hiroshi Ohata	Int J Cancer	2004	376
10	Gastric cancer epidemiology and risk factors	Douglas E Guggenheim	J Surg Oncol	2013	346

Total citations: Total citations of publications; TC: Total citations.

Table 6 The top 10 documents in co-citation analysis of publications in risk factors for gastric cancer research.

Rank	Title	First author	Source	Publication year	TC
1	Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries	Freddie Bray	Ca-cancer J Clin	2018	251
2	The two histological main types of gastric carcinoma: diffuse and so-called intestinal-type carcinoma. An attempt at a histo-clinical classification.	P Lauren	Acta Pathol Mic Sc	1965	224
3	Helicobacter pylori infection and the development of gastric cancer	N Uemura	New Engl J Med	2001	207
4	Global cancer statistics, 2002	D Max Parkin	Ca-cancer J Clin	2005	196
5	Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994	M F Dixon	Am J Surg Pathol	1996	180
6	Bias in meta-analysis detected by a simple, graphical test	M Egger	BMJ	1997	159
7	Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012	Jacques Ferlay	Int J Cancer	2015	150
8	Gastric cancer and Helicobacter pylori: a combined analysis of 12 case control studies nested within prospective cohorts	Helicobacter and Cancer Collaborative Group	Gut	2001	144
9	Interleukin-1 polymorphisms associated with increased risk of gastric cancer	M Egger	BMJ	1997	137
10	Meta-analysis in clinical trials	R DerSimonian	Control Clin Trials	1986	126

Total citations: Total citations of publications; TC: Total citations.

Keywords analysis

The list of key risk factors for GC: Keywords are high-level summaries and condensations of topics in an article[33]. High-frequency keywords represent hot topics in a research field[34]. In this study, the frequency of research on the risk factors in the keywords represents their visibility in the field. It also reflects the importance of this risk factor in the development of GC. A total of 6412 keywords were included in this study, and the top 20 risk factors in terms of study frequency were further extracted to construct a list of key risk factors for GC by excluding the subject terms (Table 7). Among them, "H. pylori infection" (n = 717), "polymorphism" (n = 326), "smoking" (n = 218), and "diet" (n = 137) were the main risk factors for GC.

Table 7 The top 20 risk factors for gastric cancer.

Rank	Keyword	TP	Rank	Keyword	TP
1	Helicobacter pylori infection	717	11	Nutrient intake	62
2	Polymorphism	326	12	DNA methylation	40
3	Smoking	218	13	Life-style	36
4	Diet	137	14	Fruit	33
5	Alcohol	112	15	Pepsinogen	32
6	IM	99	16	Promoter polymorphism	31
7	Inflammation	98	17	Lymphadenectomy	29
8	Obesity	91	18	Necrosis-factor-alpha	29
9	Caga	63	19	s-1	29
10	Biomarker	62	20	p53	27

Total publications: Total number of publications during the observation period; TP: Total publications.

Research hotspots (clustering analysis of keywords): In the keyword co-occurrence network (Figure 6), the keywords can be divided into 4 clusters according to their colors, which are: Cluster 1 (red): molecular epidemiology, involving 27 keywords including "apoptosis", "DNA methylation", "genetic polymorphisms", "necrosis-factor-aloha" and 27 other keywords; Cluster 2 (green): lifestyle, including "alcohol", "cigarette-smoking", "diet", and "obesity"; Cluster 3 (blue): postoperative risk factors, involving "chemotherapy", "endoscopic resection", "gastrectomy" and 17 other keywords; Cluster 4 (yellow): GC-related disease screening, containing only "chronic atrophic gastritis", "endoscopy", "helicobacter-pylori infection" and 11 other keywords.

Figure 6 Co-occurrence network visualization map of keywords.

Research fronts (keyword burst analysis): Keyword burst analysis can determine the evolution of research hotspots by analyzing the temporal trends of research hotspots[35]. Figure 7 shows the top 25 keywords in terms of burst intensity between 2000-2022, where the strongest keyword is gastrectomy (16.34) and the keywords with the longest burst duration are diet (2000-2009), lung cancer (2004-2013), genetic polymorphisms (2005-2014), and promoter (2005-2014). Notably, the keywords that continue to explode until 2022 are: gastrectomy (2018-2022), and model (2018-2022), respectively, which are the current issues at the forefront in risk factors for GC.

Figure 7 Top 25 keywords with the strongest citation bursts.

DISCUSSION

In this study, we used bibliometric methods to count 2514 risk factors for GC-related articles in the WOSCC database, analyze the contributions of countries, institutions, authors, and journals, and summarize hot topics and predict future research directions. To the best of our knowledge, this study is the first bibliometric analysis of publications on risk factors for GC.

General information on risk factors for GC literature

The number of annual publications has increased each year since 2000, with 72 countries, 489 journals, 2507 institutions, and 12569 authors participating in this field of study. This indicates that risk factors for GC is a popular research area at present, and it is foreseeable that more countries and researchers will participate in risk factors for GC research in the future. China, Japan, and the United States are the main contributors to publications in the field, and the United States is an important initiator and promoter of international cooperation, playing an active role in exchanging academic results and sharing academic resources. Asian institutions, such as the National Cancer Center (Goyang-si, Korea) and Nanjing Medical University (Nanjing, China), are the primary contributors in the field of risk factors for GC. Leading researchers such as Il Ju Choi, Shoichiro Tsugane, Jeongseon Kim, and Yuan Yuan have made substantial contributions. International Journal of Cancer, Gastric Cancer, Plos One, and World Journal of Gastroenterology are the four authoritative academic journals in the field of risk factors for GC, which is valuable for understanding the significant academic perspectives.

Hotspots and frontiers of risk factors for GC research

Through analyses of high-frequency keywords, highly cited literature, keyword clustering, and burst words, we identified four research hotspots: (1) H. pylori infection; (2) single nucleotide polymorphism; (3) bio-diagnostic markers; and (4) GC risk prediction models.

H. pylori infection

Keyword analysis identified H. pylori infection as the most significant risk factor for GC. Approximately 90% of GC are associated with H. pylori infection. Eradication of H. pylori reduces the inflammatory response, promotes ulcer healing, and reduces the incidence of GC. It was reported that if 30% of GC could be prevented by eradicating H. pylori, then it would be a cost effective cancer prevention strategy[36]. Since 1994, when H. pylori was recognized as a Group I carcinogen by the WHO, research on the pathogenicity, diagnosis, and treatment of H. pylori has been the focus, with new research constantly emerging. Plummer et al[37] found that H. pylori has a higher oncogenic effect than infectious agents such as hepatitis B virus and human papillomavirus, and is by far the most common oncogenic infectious agent. Currently, however, the treatment of H. pylori is facing serious difficulties due to increased antibiotic resistance, and eradication rates of less than 80% in some countries[38]. It is generally accepted that the emergence of antibiotic resistance is associated with mutations in the H. pylori gene[39]. For example, the 23S ribosomal RNA A2143G mutation is associated with clarithromycin resistance[40], the 87^th mutation in gyrA is associated with levofloxacin resistance[41], and the tet-1 mutation in 16S rRNA is associated with tetracycline resistance[42]. Currently, the Maastricht guidelines recommend the use of bismuth-containing quadruple therapy as the first-line treatment for H. pylori eradication in areas with high clarithromycin resistance[43]. Xie et al[44] have developed an H. pylori vaccine using chitosan as an adjuvant and found that it was effective in preventing H. pylori infection and inducing Th1 and Th2 immune responses. Currently, vaccine development and application remain challenging due to the weak immune response stimulated by protein antigens on the mucosal surface, the potential toxicity of exotoxin-based mucosal adjuvants, and the limited capacity to control H. pylori infection[45]. Gao et al[46] concluded that risk stratification based solely on the presence or absence of H. pylori infection is of limited value and that serologic markers based on reactions to virulence factors are more necessary. The detection of 15 H. pylori protein serum markers revealed that CagA and GroEL were independent risk predictors for the occurrence of GC.

Single-nucleotide polymorphism

Keyword analysis identified single-nucleotide polymorphism (SNP) as the predominant genetic factor in GC. The occurrence of GC is the result of the interactions between multiple genetic, environmental, and social factors[47]. Only a small percentage of these individuals exposed to risk factors such as H. pylori develop GC as the genetic risk is low[48], and there are significant differences in genetic susceptibility between individuals[49]. Additionally, studies have shown that up to 10% of GC cases are due to germline pathogenic variants caused by underlying genetic susceptibility[50]. Thus, novel and effective methods for identifying susceptibility genetic factors are expected to provide new insights into the underlying molecular pathways of tumorigenesis[51]. In recent years, genomic analysis of gastric tumors has highlighted the importance of their genetic heterogeneity. Over the last decade, SNP analysis has been widely used to screen candidate genes and detect a variety of complex human diseases, providing an approach to identify genetic loci associated with cancer heterogeneity[52]. The available evidence suggests that most SNPs are functionally neutral, yet there are still some SNPs shown to be associated with GC development and prognosis[53]. Studies have shown that SP4 gene polymorphisms (rs39302, rs7811417)[54], GBAP1 gene polymorphisms (rs140081212, rs1057941 and rs2990220)[55], and Cathepsin B gene polymorphisms (rs9009, rs6731 and rs17814426)[56] are associated with a reduced risk of GC and could provide a protective effect in its development. Although the genetic causes of cancer have been extensively studied, it is becoming increasingly clear that a large proportion of cancer susceptibility cannot be attributed to variations in protein-coding sequences, with one study specifying that more than 80% of cancer-associated SNPs occur in non-coding regions of the genome[57]. Long noncoding RNAs (lncRNA) are universally transcribed non-protein-coding transcripts in the genome that are longer than 200 nucleotides[58]. These lncRNAs play an important role and available evidence has demonstrated that many SNPs in lncRNAs lead to the modification of their interaction partners by interfering with the splicing process and the stability of mRNA conformation[59]. This in turn affects the molecular functions of cells or of tissue development and differentiation, which consequently have an impact on the expression of tumor characteristics[60] and increasing cancer susceptibility[57]. Ge et al[61] performed a clinical study using TaqMan technology on 1536 subjects evaluating three lncRNA PTENP1 genotypes. They found that subjects with the rs7853346 G allele had significantly lower risk of GC compared to those with the C allele. However, there are also some issues that must be addressed in the current research: (1) The relatively small sample size used for stratified analysis may reduce statistical power[62]; (2) limited genotypes were selected and the overall representativeness of the SNP characteristics is questionable[63]; and (3) most studies failed to consider the influence of environmental factors on the results[64].

Bio-diagnostic markers

Bio-diagnostic markers are the most promising diagnostic modality for GC, as identified through keyword analysis. Gastrointestinal endoscopy combined with gastric mucosal biopsy is currently the gold standard for the diagnosis of GC, and its sensitivity and specificity can reach 69% and 96%, respectively[65]. Nevertheless, its long operation time, the invasiveness of the operation, and the partially subjective nature of the diagnosis[66] are still issues that cannot be ignored. The American Society for Gastrointestinal Endoscopy explicitly recommends endoscopic surveillance only for patients with a specific ethnicity and family history[67]. A convenient, noninvasive way to perform mass screening is therefore necessary[68]. Bio-diagnostic markers are the most promising diagnostic modalities currently under investigation. Despite the widespread clinical use of bio-diagnostic markers such as CEA, CA19-9, CA72-4, and CA125, they often failed to meet actual clinical needs due to insufficient sensitivity and specificity[69]. Consequently, the discovery of new bio-diagnostic markers remains a pressing concern to improve the overall survival of GC patients[70]. It has been shown that Apolipoprotein C1 (APOC1)[71], interferon gamma receptor 1 (IFNGR1)[72], and hioredoxin reductase 1 (TXNRD1)[73] differ in serum concentrations between healthy and diseased populations, and reflect, to some extent, the staging, classification and progression of the disease. Thus, APOC1, IFNGR1, and TXNRD1 are potential serum biomarkers for the diagnosis of GC. Meanwhile, there is growing evidence that miRNA dysregulation is closely associated with the onset, progression and metastasis of many cancers[74]. MicroRNAs (miRNAs) are small non-coding RNAs composed of 20-25 nucleotides[75] that regulate protein translation[76] and are considered to be promising diagnostic biomarkers for cancer[77]. For instance, the diagnostic feasibility, sensitivity and specificity of miRNA-381[78], and miRNA-590-5p[79] have been investigated and confirmed.

GC risk prediction model

The GC risk prediction model is a recent research hotspot identified by burst word analysis. The reasons for GC risk prediction models to become a hot research topic are as follows: (1) Risk prediction models are a simple and effective method for assessing individual risk by quantifying the incidence of cancer[80]; and (2) risk prediction models can integrate multiple GC causative factors, including environmental, behavioral, biological, and genetic factors, to realize the accuracy of cancer risk assessment[81]. Table 8 summarizes the research characteristics and model features of the GC risk prediction model-related literature[82-94]. For instance, Pei et al[95] constructed a linear prediction model based on H. pylori infection, PGI, G-17, and the number of lesions that showed good predictive value in both the subsequent training cohort and the confirmatory cohort, providing a good basis for patient assessment and individualized treatment. The GC risk prediction model constructed by Iida et al[91] has good discriminative power and calibration scale. Eom et al[80] developed GC risk prediction models based on different genders and showed good accuracy and predictability. Although existing risk prediction models have the desired predictive power, it is evident that many models still have a high risk of bias and also lack validation components[96]. In addition, the lack of calibration evaluation and unclear application settings have hindered the practical application and widespread spread of predictive models[97].

Table 8 Research characteristics of gastric cancer risk prediction models.

Research characteristics				Model characteristics			Findings	Ref.
Country	Research design	Number of participants	Data types	Model types	AUC	C-index
China	Prospective cohort study	435673	①②③④	c	-	0.736	The GCRS can be an effective risk assessment tool for tailored endoscopic screening of GC in China. RESCUE, an online tool was developed to aid the use of GCRS	Zhu et al[82]
China	Retrospective study	6005	①④⑤⑥⑦	-	0.708	-	Li’s prediction model performs the best for risk stratification in the screening, detection, and diagnosis of GC and precancerous lesions, whereas the overall performance of the other three models is similar	Hu et al[83]
South Korea	Retrospective cohort study	1157	①④⑤⑦	a	0.894	-	The 4-point discriminative model may help identify patients with a normal serological test who are nonetheless at risk of developing GC	Cho et al[84]
China	Cohort study	89568	①②④⑨	b	0.97	-	This model could enable a potentially more cost- effective endoscopic surveillance program, as well as to exclude very low-risk patients from unnecessary surveillance	Leung et al[85]
China	Retrospective cohort study	2287	①②④⑩	a	0.684	-	The present study established a predictive model to assess the risk of GC using high-evidence genetic variants and detected the potential gene-environment interaction, which may be helpful in prevention of the cancer	Qiu et al[86]
China	Case-control study	383	①②③④⑤	a	0.883	-	This model is simple, convenient, and economical, has good patient compliance, is easy to implement clinically, is easy to concentrate medical resources, and is expected to identify high-risk groups at an early stage, then to increase the detection rate of GC	Tao et al[87]
America	Case-control study	140	①②③④⑦	a	0.9495	-	The addition of ethnic and cultural variables, particularly the immigration/generation, to conventional risk factor variables improved the ability of models to identify individuals at high risk for GC	In et al[88]
Japan	Case-control study	1431	①④⑦⑧	b	0.899	-	XGBoost outperformed logistic regression and showed the highest AUC value	Taninaga et al[89]
China	Cross-sectional study	14929	①②③④⑤⑥⑦	a	0.76	-	The prediction rule had good performance and showed significantly better discrimination ability to identify a patient with GC than three other alternative prediction methods	Cai et al[90]
Japan	Cohort study	5648	①②⑦⑧	c	0.790	-	We developed a risk assessment tool for gastric cancer that provides a useful guide for stratifying an individual’s risk of future gastric cancer	Iida et al[91]
China	Cross-sectional study	12112	①②③④⑤⑥⑦	c	0.811	-	A serological biopsy composed of the five stomach-specific circulating biomarkers could be used to identify high-risk individuals for further diagnostic gastroscopy, and to stratify individuals’ risk of developing GC and thus to guide targeted screening and precision prevention	Tu et al[92]
Japan	Prospective cohort study	19028	①②③④⑤⑦	c	-	0.777	In this study, the authors developed a model and a simple scoring system to estimate an individual's risk of developing GC, based on factors such as H. pylori antibodies, serum pepsinogen levels, and lifestyle habits	Charvat et al[93]
South Korea	Case-control study	217	①②③④	a	0.888	-	This study provides the first predictive model for assessing the risk factors for GC in Korea, where the incidence rate of GC is high. This study has also identified new risk factors for GC, such as drinking tap water	Lee et al[94]

^aLogistic regression model.

^bMachine learning algorithm

^cCox proportional hazards model.

① demographic data; ② personal history; ③ family history; ④ gastroscopy; ⑤ pepsinogen; ⑥ gastrin-17; ⑦ Helicobacter pylori infection status; ⑧ HbA1c; ⑨ medication history; ⑩ single nucleotide polymorphisms.

XGBoost: Extreme gradient boosting; GC: Gastric cancer; AUC: Area under the curve; RESCUE: Risk Evaluation for Stomach Cancer by Yourself; GCRS: GC risk score; H. pylori: Helicobacter pylori.

Analysis of risk factors for GC

The list of key risk factors for GC: In this study, we extracted the top 20 risk factors in terms of study frequency and constructed a list of key risk factors for GC based on the frequency magnitudes. The ranking order represents the level of interest in the risk factors for GC field for each risk factor, and, to some extent, reflects its importance in the development of GC. Based on our analysis, we believe that H. pylori infection, SNP, smoking, diet, and alcohol are the most important risk factors for the development of GC. Previous studies also support our findings, for example, H. pylori is thought to cause 65% to 80% of GC cases[98], and smoking contributed to approximatley 11% of GC cases[99]. Currently, GC prevention and treatment strategies have a few limitations, such as the propagation of one single GC prevention and treatment strategy[100]; the reliance on endoscopy, which has difficulty in population dissemination[101]; and the inclusion of a large number of risk factors without prioritization[102]. To dress these issues, prevention and screening efforts should center on a clear list of key risk factors, and the list of risk factors derived from this study may potentially contribute to it. Accordingly, it is important to first control H. pylori infection while avoiding smoking, diet and alcohol consumption habits. Meanwhile, genotyping associated with GC could be screened by SNP microarray detection technology[103], which could effectively determine GC occurrence.

Common risk factors for different tumors: In this study, analysis was conducted to explore the common risk factors of GC with other tumors. For example, Chen et al[104] found an association between the MDM2 promoter rs937283 and lung cancer risk, as well as an association with GC risk in men, population with smoking and alcohol consumption characteristics. Lee et al[105] found that proton pump inhibitor use for ≥ 2 years was not associated with increased risk of gastrointestinal cancers, including gastric and pancreatic cancers. Bidel et al[106] also have not yet found an association between coffee consumption and risk of gastric and/or pancreatic cancer. The exploration of common risk factors can help optimize tumor prevention and treatment strategies and reduce social public health costs.

Gender differences in risk factors for GC: In the study, we also found that there were some gender differences in risk factors for GC. For example, one study found that the prevalence of H. pylori infection is greatly increased in adult males, and this gender difference is one of the factors contributing to the emergence of a male predominance in diseases associated with H. pylori infection, including GC[107]. The same finding was observed by other studies, including Song et al[108] and Verma et al[109]. This may be a key factor contributing to gender differences in the incidence of GC[110] and is closely associated with gender differences in clinical staging[111] and pathological features[112].

To the best of our knowledge, this is the first bibliometric study for risk factors for GC. This study has helped to identify frontier and hot topics and to select appropriate research directions, collaborating institutions, and publishing journals. The key risk factors for GC may provide a reference for orderly clinical intervention. This study also has the following limitations: (1) The literature included in this study is in English, and articles published in other languages may not be included; (2) this study only searches the WOS database, and important literature included in other databases may be missed; and (3) this study's literature search only extends to December 04, 2022, and as some articles are published online first and not yet actually published, there is a certain risk of bias.

CONCLUSION

The number of publications in the risk factors for GC field has grown steadily over the past two decades. H. pylori infection was the most important risk factor for GC; SNP is the most dominant genetic factor in GC, and bio-diagnostic markers are the most promising diagnostic modality for GC; GC risk prediction model is the latest research hotspot. A list of key risk factors for GC was also constructed, and we identified H. pylori infection, SNP, smoking, diet, and alcohol as the most important risk factors for the development of GC, and proposed corresponding strategies for GC prevention.

ARTICLE HIGHLIGHTS

Research background

The absence of specific symptoms in patients with gastric cancer (GC) causes great difficulties in early diagnosis of the disease, therefore, it is very valuable to clarify the risk factors for early diagnosis and treatment of GC and to improve the survival of GC.

Research motivation

Researchers have conducted a large number of studies on risk factors in GC. However, a summary of the current state of research is lacking.

Research objectives

This study was conducted to enable physicians to keep abreast of the latest research evidence in the clinical field and to make increasingly informed decisions.

Research methods

The articles covered in this study were obtained from the Web of Science database and the data were processed through CiteSpace and VOSviewer software.

Research results

A total of 2514 papers from 72 countries and 2507 research institutions were retrieved. China, National Cancer Center, and Shoichiro Tsugane were the most productive country, institution, or author, respectively. The research hotspots in the study of risk factors for GC are summarized in four areas, namely: Helicobacter pylori (H. pylori) infection, single nucleotide polymorphism, bio-diagnostic markers, and GC risk prediction models.

Research conclusions

A list of key risk factors for GC was also constructed, and we identified H. pylori infection, single-nucleotide polymorphism, smoking, diet, and alcohol as the most important risk factors for the development of GC, and proposed corresponding strategies for GC prevention.

Research perspectives

GC risk prediction modeling is the latest research hotspot and will likely be the future direction of research in the field.