Cancer is not only a major public health problem worldwide, but also one of the leading causes of human mortality. Host-microbe immune interactions profoundly influence cancer development, progression and treatment outcomes (1–11). Fungi and bacteria co-colonize the mammalian skin epithelium, respiratory tract, gastrointestinal tract and reproductive organs, forming a complex ecosystem of microbe-microbe and host-microbe interactions that significantly impact human health (12–21). Although only ~0.1% of microbial DNA is present in the gut (22), fungal infections cause more than 1.5 million deaths worldwide each year (23).

There is growing evidence linking the human microbiome (bacteria, fungi and viruses) to cancer and cancer outcomes (24,25). In recent years, several bacteria have been observed to be associated with cancer development and progression. Helicobacter pylori infection is the strongest risk factor for the development of malignant tumors in the stomach, and epidemiologic studies have determined that the attributable risk of gastric cancer due to Helicobacter pylori is ~75% (26). At the same time, in the lower gastrointestinal tract, genotoxic Escherichia coli, Bacteroides fragilis, Streptococcus bovis and Fusobacterium nucleatum are associated with the pathogenesis of colorectal cancer (27). Among these cancer-associated bacteria, they can modulate host immunity and cause chronic inflammation, which is thought to have oncogenic effects. Recent reports have also shown that bacterial DNA circulating in the blood of cancer patients can be used as a predictive biomarker for tumors (28,29) and intracellular bacteria have been found in numerous tumor types (30).

In recent times, scientists have been exploring the link between cancer and bacteria and viruses, but few studies have focused on the relationship between fungi and cancer. Although the human fungal biome is less abundant than the bacterial group, it can still significantly affect human health (31–33). Previous studies have demonstrated that fungal DNA is present in most human cancer tissues. However, the role and impact of the tumor mycobiome on cancer pathogenesis remain largely unknown.

Narunsky-Haziza et al (34) statistically characterized the tumor mycobiome in 17,401 tissue, blood and plasma samples from four independent cohorts of 35 cancer types using internal transcribed spacer sequencing and whole-genome sequencing methods. Statistics revealed that 5 fungi, including Malasseziomycetes, Saccharomycetes, Diothideomycetes, Sordariomycetes and Candida, were significantly enriched in all types of cancer (lung, breast, melanoma, osteosarcoma, ovarian, gastric, colorectal and head and neck tumors). By contrast, certain specific fungi (Microbotryomycetes, Wallemiomycetes, Agaricomycetes, Tremellomycetes) were only found in specific cancer sites, so the tumor mycobiome is prevalent in most cancers in humans (35).

Fungal biome imbalance was found in patients with gastric cancer. The fungal biome characteristics of patients with gastric cancer differed significantly from controls, with reduced diversity and enrichment of Candida and Alternaria in gastric cancer tissues (36). Colorectal fungi were altered in patients with colorectal cancer compared to normal subjects. Patients with colorectal cancer had increased Malasseziomycetes, decreased Saccharomycetes and a distorted ratio of Basidiomycota to Ascomycota (37). The distribution of the fungal genera Aspergillus, Malassezia, Rhodotorula, Pseudogymnoascus, Kwoniella, Talaromyces, Debaryomyces, Moniliophthora, Pneumocystis and Nosemia was altered in colorectal cancer, as verified in independent Chinese and European cohorts (38).

Analysis of the tumor mycobiome in various body parts showed a maximum of 1 fungal cell per 10,000 tumor cells (39). In The Cancer Genome Atlas (TCGA) primary tumors, the average relative abundance of bacteria and fungi was 96 and 4%, respectively, which confirms the lower abundance of fungi compared to bacteria (34). It has been found that there is a symbiotic rather than competitive ecological interaction between fungal and bacterial biomes in the tumor microenvironment (35). However, this differs from the manifestation of alternating fungal and bacterial populations in the gastrointestinal tract (9,40).

The tumor mycobiome not only resides in tumors but also promotes tumor progression and metastasis and spreads systemically by affecting the human immune system, maintaining a pro-inflammatory environment, producing aflatoxins, attenuating cell adhesion mechanisms and fungal-bacterial interactions (Fig. 1).

Numerous studies have indicated the potential of bacteria as biomarkers for the diagnosis of colorectal cancer (64). Wirbel et al (80) and Thomas et al (81) performed meta-analyses to identify several bacteria enriched in colorectal cancer with utility as diagnostic biomarkers for colorectal cancer.

By contrast, with the study of the tumor microbiome, fungi, in addition to bacteria, may be used as biomarkers for the noninvasive diagnosis of tumor patients. In TCGA cohorts, fungal biome richness varied significantly among cancers (35). For instance, Candida, Saccharomyces cerevisiae and Cyberlindnera jadinii are highly abundant in the gastrointestinal tumor fungal organism community and Blastomyces and Malassezia are highly abundant in lung cancer and breast cancer, respectively (39).

A further study performed qualitative and quantitative analyses of 20 different fungal DNAs released into the bloodstream and the results suggested that they may be used to distinguish patients with cancer from healthy individuals, even in early-stage disease (35). This suggests that the tumor fungal biome has utility in cancer diagnosis.

In the study conducted by Lin et al (64), the average areas under the receiver operating characteristic curves (AUC) of pure bacterial biomarkers was only 73%, while the combination of fungal and bacterial mixed biomarkers showed a significant improvement in diagnostic performance with an average AUC of 83% and an increase in the relative change in AUC of 1.44–10.60% (64). This suggests that the combination of fungal and bacterial biomarkers is more accurate than the combination of pure bacterial species in differentiating patients with colorectal cancer from healthy individuals, thus highlighting the potential use of the tumor mycobiome in clinical diagnostic applications.

In the study of colorectal cancer conducted by Liu et al (77), a comprehensive analysis of different national microbiomes was performed using colorectal cancer metagenomic datasets of 8 different cohorts. They found that fungi, archaea and viruses were able to distinguish patients with colorectal cancer from healthy controls in multiple geographic cohorts (77). Coker et al (38) successfully distinguished 184 patients with colorectal cancer from 204 healthy controls by detecting fungal biomes in the stool.

Candida is transcriptionally active in gastrointestinal tumors (35). Enrichment of tumor-associated Candida DNA was found to be significantly associated with reduced survival in patients with gastrointestinal tumors due to the association of Candida with gene expression for cytosolic DNA sensing, Toll-like receptor signaling and Nod-like receptor signaling in gastric cancer (39). This not only suggests that Candida increases the severity of gastrointestinal tumors but also that Candida may be a promising biomarker for predicting disease outcomes.

The use of antimicrobial agents that target known pathogenic microorganisms effectively prevents the onset and progression of the disease. The use of targeted antifungal agents is helpful in the prevention or treatment of gastrointestinal cancers (37). In a mouse model of human pancreatic ductal adenocarcinoma, Malassezia infiltrates and accelerates the progression of human pancreatic ductal adenocarcinoma, a condition that can be reversed by antifungal treatment (31). Antifungal treatment targeting Malassezia resulted in a 40% reduction in the incidence of pancreatic cancer in mice (31). In a mouse model of esophageal cancer, the oral fungus Dictyostelium significantly increased the severity of esophageal squamous cell carcinoma, a condition that could be reversed by antifungal treatment (82).

In addition, the use of fungal probiotics can prevent and treat gastrointestinal cancers. Probiotics are microorganisms that improve health when consumed in the correct amounts. The most common probiotics are bacteria that have been shown to inhibit the proliferation of pathogenic intestinal microorganisms and to prevent carcinogenic inflammation in the esophagus, stomach, pancreas and colorectum (83–85). By contrast, fungi can also be ingested as probiotics and have been reported to be used to alleviate gastrointestinal cancers (37). The Helicobacter pylori eradication rate improved when Saccharomyces boulardii was combined with Lactobacillus gasseri, Lactobacillus reuteri, Lactobacillus acidophilus, Streptococcus faecalis, Bacillus subtilis and Bifidobacterium (86).

The use of fungal probiotics in treating gastrointestinal cancers is of great importance. Because of their characteristic cellular structure, they can survive in the unfavorable environment of the gastrointestinal tract (87,88). As more research on the efficacy and safety of fungal probiotics is conducted, they may directly or indirectly modulate the tumor microbiome to prevent and treat gastrointestinal cancers (37).

With the above summary and analysis, the fungal biome that targets and attacks tumors is likely to be a good way to treat cancer.

This review provides a comprehensive summary of the impact of the tumor mycobiome on cancer pathogenesis. The tumor mycobiome promotes tumor progression and metastasis by affecting the human immune system, maintaining a pro-inflammatory environment, producing aflatoxins, attenuating cellular adhesion mechanisms and fungal-bacterial interactions. Furthermore, the tumor mycobiome also has tremendous potential for cancer prevention, diagnosis and treatment.

Although studies on the effect of the tumor mycobiome on cancer pathogenesis have become more frequent, the results are not uniform because of the differences between different populations and inconsistent standards for metagenomic data generation and processing. The future development of standardized and low-cost sequencing technologies and pipeline analysis methods to improve the quality of data collection and analytical processing, and the initiation of longitudinal studies with large sample sizes in different populations to clarify the specific mechanistic relationships, will be crucial for research in this field.