High-cited articles are one of the most valuable indicators in bibliometrics with extremely high academic importance in a field. Table 5 lists the top 20 high-cited papers in original research (published between 2012 and 2020). Based on the subject matter, we summarized the following three points:

First, there are differences in IM between patients with CRC and healthy controls (Chen et al., 2012; Wang T. et al., 2012; Ahn et al., 2013; Weir et al., 2013). Microbiome transformation may occur in the early stages of CRC and fecal metagenomes may reveal microbial characteristics specific to CRC (Yachida et al., 2019). Tumor and non-tumor related microbiota in CRC may be different; mucosal microbiota is only partially reflected in fecal microbiota, and CRC can be stratified based on the higher-level mucosal microbiota co-abundance group (Flemer et al., 2017). Furthermore, IM can be used for CRC screening. Metagenomic analysis of GM can be used as a tool for targeting noninvasive biomarkers to diagnose CRC (Yu J. et al., 2017). The sensitivity of CRC diagnosis can be increased by combining IM with the standard fecal occult blood test (FOBT) (Zeller et al., 2014). Second, IM may play a key role in tumorigenesis, inflammation-cancer, and adenoma-carcinoma transition. for example, laboratory-type mice reconstructed with IM from wild-type mice exhibited improved resistance to colorectal tumorigenesis (Rosshart et al., 2017). Moreover, germ-free mice colonized with IM from tumor-bearing mice had a relatively higher population abundance related to inflammation-driven tumor formation (Zackular et al., 2013). NOD₂-mediated ecological imbalance made mice susceptible to colitis and CRC (Couturier-Maillard et al., 2013), whereas activating Gpr109a (the receptor of the symbiotic metabolite butyrate) inhibited colitis and carcinogenesis (Singh et al., 2014). Moreover, inflammation can promote CRC, and IM has been identified as an inflammatory target that affects CRC progression (Arthur et al., 2012). IM can also evolve along the colorectal adenoma-carcinoma sequence (Feng et al., 2015). Patients with familial adenomatous polyposis carry colonic biofilms containing carcinogenic bacteria, and tumor-prone mice colonized with carcinogenic bacteria show a faster tumor onset and higher mortality (Dejea et al., 2018). Third, Fusobacterium nucleatum (F. nucleatum) is not only a key pathogenic factor, but also a biomarker for the therapeutic effect of CRC. F. nucleatum infection is prevalent in human CRC (Castellarin et al., 2012) and may potentiate intestinal tumorigenesis, regulate the tumor microenvironment (Kostic et al., 2013), and promote chemoresistance to CRC by regulating autophagy (Yu T. et al., 2017). Moreover, the amounts of F. nucleatum and the IM-produced cytokines IL-12 and TGF-β have been shown to be different, which led to different proportions of lymphocytes in different CRCs (Saito et al., 2016). The amount of F. nucleatum in CRC tissues is also related to shorter survival and may act as a prognostic biomarker (Mima et al., 2016).

Table 6 shows the top ten high-cited reviews (issued between 2012 and 2019), nearly half were from Nature Reviews Gastroenterology & Hepatology (n = 2) and Nature Reviews Microbiology (n = 2). Two review articles (Marchesi et al., 2016; Kho and Lal, 2018) outlined the key role of IM in host health and disease. Several review articles (Louis et al., 2014; Gagnière et al., 2016; O'Keefe, 2016; Tilg et al., 2018) outlined the links between diet, IM and metabolites, and CRC. Furthermore, some reviews detailed the mechanisms of inflammation-driven IM dysbiosis (Zeng et al., 2017), outlined the interplay of bile acids and microbiota in gastroenteritis inflammation and carcinogenesis (Jia et al., 2018), mentioned the theoretical hypothesis of the “driver-passenger” model, arguing that some microbes may cause adenomas and cancers (Tjalsma et al., 2012), and summarized the impact of IM on tryptophan metabolism-mediated intestinal immunity (Gao et al., 2018).

Considering that some classical papers (particularly before 2012) still had important significance, we searched for the most cited references to find important papers that may have been ignored. Figure 6A and Figure 6B show the top 20 most-cited references and their citation relationships. In 2009, Wu et al. (2009) demonstrated that human colon bacteria could promote colon tumorigenesis by activating the T-assisted type 17 T cell response. In 2010, Caporaso et al. (2010) developed QIIME (a tool for unpacking massive high-throughput sequencing data). A human intestinal metagenomic study (Qin et al., 2010) based on metagenomic sequencing has provided a broad perspective on the important functions of intestinal bacteria. These technologies have increased support for IM/CRC research. In 2011, Sobhani et al. (2011) applied pyrosequencing technology to report that colon cancer is associated with microbial dysbiosis, opening up a new field for CRC screening and pathophysiology research. Marchesi et al. (2011) compared the microbial composition between colon tumors and the adjacent non-malignant colonic mucosa, revealing significant differences in the IM of the two sites.

Several classic papers were identified through historiographic analysis ( Figure 7 ). Some papers had been mentioned in the part of most-cited articles and reviews. Moreover, Wu et al. (2013) showed that the IM of CRC was characterized by the enrichment of potential pathogens, such as Fusobacterium and Campylobacter and the reduction of butyrate-producing bacteria. Zackular et al. (2014) showed that IM can be used as a screening tool to detect precancerous lesions and cancers in CRC. Nakatsu et al. (2015) classified IM communities in the intestinal mucosa at different stages of CRC and found that with the progression of CRC along the “adenoma-cancer” sequence, mucosal microflora can establish a microecosystem. In 2014, a review (Louis et al., 2014) outlined the links between diet, metabolism, and CRC, showing that short-chain fatty acids such as acetates, propionates, and butyrate can inhibit inflammation and cancer, while some microbial metabolites (such as secondary bile acids) can promote carcinogenesis. In 2016, a review (O'Keefe, 2016) summarized the links between diet, microorganisms and their metabolites, and colon cancer, showing that meat increases the risk of colon cancer, but foods rich in fiber inhibit the risk, which may be related to IM. A review (Gagnière et al., 2016) discussed the relationship between IM and CRC, with an emphasis on dysbacteriosis and potential characteristics of carcinogenic bacteria, such as genotoxicity and other virulence factors, inflammation, host defense regulation, bacterial metabolism, oxidative stress, and antioxidant defense regulation.

To identify the hot topics and central issues in IM/CRC research, it is necessary to examine key index-keywords (Yang S. et al., 2022; Zyoud et al., 2022). In this research, a total of 6851 keywords included 3144 author keywords and 3707 keywords plus were acquired from publications.

Highly frequent author keywords (remove main search terms) included “inflammation”, “probiotics”, “inflammatory bowel disease”, “diet”, “butyrate”, “Fusobacterium nucleatum”, “chemotherapy”, “ulcerative colitis”, “apoptosis”, “ulcerative colitis”, “prebiotics”, “short-chain fatty acids”, “carcinogenesis”, “metagenomics”, “biomarkers”, “bile acids”. Highly frequent keywords plus (remove main search terms) included “inflammation”, “Fusobacterium-nucleatum”,”chain fatty-acids”, “inflammatory-bowel-disease”, “ulcerative-colitis”, “carcinogenesis”, “tumorigenesis”, “diet”, “Escherichia-coli”, “butyrate”, “NF-kappa-B”, “Crohn’s-disease”, “dietary fiber”, “metabolism”, “probiotics”, “metabolites”, “oxidative stress”, “enterotoxigenic Bacteroides-fragilis”.

A cluster analysis of high-frequency keywords can identify hot topics in a field (Yang S. et al., 2022; Zyoud et al., 2022). A cluster analysis was conducted based on the co-occurrence of common keywords (frequency ≥ 20). Each clustered keyword unit was considered a category based on the same color ( Figure 8A ).

Evolving keywords can reflect frontier knowledge in a special field. We predicted the trends in IM/CRC research in the next few years using overlay visualization in VOSviewer. As shown in Figure 8B , more yellow nodes were found in clusters 1, 2, and 5 than that in the other clusters, and the main keywords included search terms such as “colorectal cancer”, “gut microbiota” and “gut microbiome”, and other terms such as “chemotherapy”, “immunotherapy”, “therapy”, “efficacy”, “biomarker”, “Fusobacterium nucleatum”, “metastasis”, “metabolites”, “short-chain fatty acids”, “immunity”.

Some studies (Wong et al., 2017b; Li et al., 2019) have shown that gavage of fecal samples from patients with CRC in germ-free and normal mice can promote the progression of intestinal adenoma and carcinogenesis. Germ-free mice colonized with IM from tumor-bearing mice showed increased tumorigenesis (Zackular et al., 2013), whereas laboratory mice transplanted with IM from wild mice showed increased resistance to colorectal tumorigenesis (Rosshart et al., 2017). IM depletion with antibiotics can result in a significant decrease in subcutaneous tumor and liver metastasis burdens in mice (Zackular et al., 2013; Sethi et al., 2018). Many studies (Konstantinov et al., 2013; Louis et al., 2014; Wong and Yu, 2019; Cheng et al., 2020; Janney et al., 2020) have shown that Fusobacterium nucleatum, Escherichia coli, Enterococcus faecalis, and enterotoxigenic Bacteroides fragilis are closely related to CRC tumorigenesis, whereas butyrate-producing bacteria such as Faecalibacterium, Roseburia, Clostridium and Lachnospiraceae may inhibit the onset and development of CRC.

General risk population screening can reduce the morbidity and mortality associated with CRC. Accurate, noninvasive screening tests can significantly reduce the global health burden of CRC. Multiple studies (Chen et al., 2012; Wang T. et al., 2012; Ahn et al., 2013; Weir et al., 2013; Wu et al., 2013; Zeller et al., 2014; Flemer et al., 2017) have shown that the IM of patients with CRC was different from that of patients without CRC. IM can be used as a novel biomarker for the non-invasive diagnosis of CRC (Zackular et al., 2014; Liang et al., 2017; Wong and Yu, 2019), and metagenomic analysis of IM provides a rich source for CRC screening.

In 2017, Yu et al. (2017). found that 20 gene markers were differentially expressed in CRC and control samples, among which butyryl-coenzyme A dehydrogenase from F. nucleatum and RNA polymerase subunit from Micromonas β showed good diagnostic value, with an area under curve (AUC) of 0.84. In 2018, Dai et al. (2018). analyzed metagenomic data from patients with CRC and identified seven species, including Bacteroides fragilis and F. nucleatum enriched in CRC as potential diagnostic markers that could be used in different populations to distinguish CRC patients from healthy controls (AUC = 0.80). In 2019, Nature Medicine published three articles in succession (Thomas et al., 2019; Wirbel et al., 2019; Yachida et al., 2019). Yachida et al. (2019) found that the abundance of Firmicutes, Fusobacteria and Bacteroidetes showed an upward trend with CRC progression; propionate and butyrate were the most abundant metabolites, and the model combining bacterial species, KO genes, and metabolites was the best in terms of resolution, and found a panel of 55 bacterial markers linked to CRC. Wirbel et al. (2019) carried out a meta-analysis of eight shotgun metagenomic studies of CRC and found that the abundance of 29 strains increased in patients with CRC and revealed microbial characteristics specific to CRC. Thomas et al. (2019) conducted a fecal metagenomic meta-analysis from five available datasets and two new cohorts to identify common IM characteristics across different populations of CRC, constructed a CRC disease prediction model containing 16 species (AUC > 0.8), validated it in two additional cohort datasets, and found that the choline trimethylamine lyase gene among the flora genes was enriched in CRC. In 2020, Yang et al. (2020). identified 22 microbial marker genes closely related to CRC and verified these using qPCR. Among them, the biomarker of the gene from Coprobacillus showed a high diagnostic value (AUC = 0.93).

In addition, Zeller et al. (2014) showed that combining metagenomic analysis with FOBT could increase the sensitivity of CRC detection. F. nucleatum can be used as a biomarker for early CRC screening and prognosis. Wong et al. (2017a) identified F. nucleatum as a valuable marker for improving the diagnostic performance of fecal immunochemical tests, with a complementary role in the detection of lesions. Guo et al. (2018) showed that the ratio of F. nucleatum to the probiotics Bifidobacterium and Lactobacillus is a valuable biomarker for early CRC screening. Notably, the characteristic detection of enteric viruses (Nakatsu et al., 2018) and the fungal microbiota (Coker et al., 2019) can also be used for CRC screening.

Efficacy is the most critical factor in the evaluation of antitumor treatment. Research on the impact of IM on cancer therapy is the most important area of cancer microbiome research. It has been confirmed that IM can mediate treatment outcomes of CRC (Wong and Yu, 2019).

Concretely speaking, the mechanisms of IM affect CRC involve many factors, such as pathogenic bacteria and their virulence factors, inflammation, bacterial metabolites, immunity, oxidative stress, intestinal barrier disruption, and so on (Gagnière et al., 2016; Cheng et al., 2020; Janney et al., 2020).

Firstly, we have discussed how pathogenic bacteria such as pks⁺ E. coli, E. faecalis and ETBF can induce DNA damage by inducing inflammation and oxidative stress. IM may induce the onset and development of CRC through two modes: the “Alpha-bugs” model (Sears and Pardoll, 2011) and the “Driver-passenger” model (Tjalsma et al., 2012). (1) The “Alpha-bugs” model believes that IM with unique virulence characteristics (Alpha-bugs bacteria), such as ETBF, can directly lead to intestinal epithelial cells carcinogenesis by secreting toxic proteins such as BFT, and ETBF-driven IM changes can cause an abnormal mucosal immune response and accumulation of cancerous intestinal epithelial cells. (2) The “Driver-passenger” model considers certain intestinal bacteria (such as E. faecalis, E. coli and ETBF) as drivers to induce DNA damage and promote carcinogenesis, subsequently, the inherent drivers will be replaced by some opportunistic pathogens or even beneficial bacteria-passengers such as Fusobacterium and Streptococcus gallolyticus that are more suitable for survival in the intestinal tumor microecology.

Secondly, chronic intestinal inflammation is generally regarded as a key factor for the progression of colitis-associated CRC (a subtype of CRC that develops directly from inflammatory bowel disease [IBD]), which is supported by the much higher incidence of CRC in patients with IBD, especially those with ulcerative colitis (Arthur et al., 2012; Louis et al., 2014). Microbial symbionts are the key determinants of gut inflammation. IM imbalance can cause host metabolic and immune changes, inducing chronic inflammation and leading to tumor progression (Gagnière et al., 2016; Cani and Jordan, 2018). Crosstalk between microbiota and bile acid also plays a vital role in gastrointestinal inflammation and carcinogenesis (Jia et al., 2018). In addition, the inflammatory tissue environment is conducive to the disturbance of the IM, which is usually characterized by the massive reproduction of specific bacterial species that can use more abundant nutrients in the inflammatory intestine (Zeng et al., 2017). The most cited original study (Arthur et al., 2012) in this paper showed that colitis can promote tumorigenesis by changing the composition of IM and inducing the expansion of genotoxic microorganisms. In summary, colitis-associated CRC mainly occurs through the inflammation-cancer pathway, in which IM plays an important role (Schmitt and Greten, 2021).

Thirdly, IM can affect CRC by means of metabolites, such as secondary bile acids, trimethylamine-N-oxide (TMAO), hydrogen sulfide (promote inflammation and carcinogenesis), and SCFAs such as propionate and butyrate (inhibit inflammation and cancer) (Konstantinov et al., 2013; Louis et al., 2014). Bile acids are metabolized by enzymes from IM, which play a vital role in intestinal immunity, inflammation, and tumors (Cai et al., 2022), and bile acid-microbiome crosstalk can affect gastrointestinal inflammation and carcinogenesis (Jia et al., 2018). Secondary bile acid production may be increased in CRC patients (Wirbel et al., 2019). The IM-derived metabolite, formate, may also exacerbate CRC progression (Ternes et al., 2022). High plasma TMAO (Bae et al., 2014) and hydrogen sulfide (Carbonero et al., 2012) levels are positively associated with high CRC risk. In contrast, IM can promote the excessive proliferation of MSH₂-deficient colon epithelial cells by providing carbohydrate-derived metabolites such as butyrate, thereby regulating the host immune system (Belcheva et al., 2014). Butyrate can reduce CRC cell proliferation and stimulate apoptosis (Weir et al., 2013; Singh et al., 2014; Hu et al., 2015). A meta-analysis showed that lower fecal concentrations of three major SCFAs (acetic acid, propionic acid, and butyric acid) were associated with a higher risk of CRC (Alvandi et al., 2022).

Last, the immune system mediates the effect of IM on CRC. Changes in crosstalk between the mucosal immune system and IM are considered to be the core defects leading to chronic gut inflammation and cancer progression (Liu et al., 2013). IM can promote tumor growth in mice by regulating immune responses, such as increasing interferon gamma (IFN-γ)-producing T cells and decreasing interleukin 17a (IL-17a)- and IL-10-producing T cells (Sethi et al., 2018). In addition, IM plays a role in memory T cell formation (Liu et al., 2020) and can stimulate CRC cells to produce chemokines that facilitate the recruitment of beneficial T cells to the tumor tissue (Cremonesi et al., 2018). Crosstalk between IM and monocyte-like macrophages can mediate an inflammatory response to promote colitis-related tumorigenesis (Yang et al., 2020b). IM can regulate the host immune system by regulating L-tryptophan metabolism, which plays a crucial role in the balance between intestinal immune tolerance and IM maintenance (Gao et al., 2018). Furthermore, changes in IM can lead to changes in glycerophospholipid metabolism, thereby affecting the therapeutic effect of immunotherapy (Xu et al., 2020).

IM modification in CRC management is of great significance, as it not only prevents the formation and progression of CRC but also improves the clinical efficacy of cancer patients and reduces adverse events (Fong et al., 2020; Kaźmierczak-Siedlecka et al., 2020). Currently, IM intervention to adjust CRC mainly includes the following aspects: