Intratumoral fungal burden of Candida tropicalis as a novel prognostic biomarker for recurrence and mortality in colorectal cancer.

INTRODUCTION
Colorectal cancer (CRC) is one of the most prevalent cancers worldwide, ranking third in incidence and second in mortality, with an estimated 1.9 million new cases and 904,000 cancer‐related deaths in 2022.
1
Despite dramatic progress in the diagnosis and treatment of cancer—including surgery, chemotherapy, and immunotherapy—the incidence and mortality rates of CRC have continued to increase, and its burden is projected to increase to 3.2 million new cases and 1.6 million deaths by 2040.
2
Elucidating the mechanisms involved in carcinogenesis, cancer progression, and metastasis is essential for improving prognosis in patients with CRC. Biomarkers are crucial tools for the early detection of cancer, as well as for predicting cancer progression and therapeutic response.
3
,
4
,
5
Therefore, over the past three decades, various research efforts have been undertaken, with significant advances being made in genomics and molecular biology. However, few biomarkers—such as KRAS gene mutations, BRAF gene mutations, and microsatellite instability status—have been translated into clinical practice.
6
,
7
,
8
,
9
,
10
This highlights the immense clinical need for further investigation of novel prognostic and diagnostic biomarkers based on the discovery of biological mechanisms.
Recent studies have investigated the association between the human microbiome and CRC, and identified potential carcinogenic bacteria such as Bacteroides fragilis, Escherichia coli, and Fusobacterium nucleatum.
11
,
12
,
13
,
14
However, the role of gut and intratumoral fungi (also known as the mycobiome) in carcinogenesis and the progression of CRC remains underexplored because of their low abundance in the human gut; fungi constitute 0.01% to 0.1% of the human gut microbiome.
15
,
16
,
17
Fungi are microeukaryotes that exist in various anatomic sites of the human body with a stable colonization. Despite their relatively small proportion in the human microbiome, they significantly influence host health, including immunity, metabolism, and organ nutrition.
18
,
19
Furthermore, recent evidence suggests that fungi significantly influence a variety of diseases, including cancers.
16
,
20
,
21
Because of significant specific mycobial dysbiosis in patients with CRC, various studies have attempted to elucidate the molecular mechanism of carcinogenesis and cancer progression promoted by fungi.
22
,
23
,
24

Candida tropicalis is a type of commensal fungus that exists in the human body, particularly in the skin, gastrointestinal tract, and urinary tract.
25
,
26
Typically, this fungus is nonpathogenic in healthy individuals; however, it can cause invasive infections in immunocompromised hosts. As for the association with CRC, C. tropicalis is reported to be abundant in CRC tissues; moreover, this fungus burden was also significantly increased in stool samples from patients with CRC when compared with those from healthy individuals.
20
,
27
In addition, several recent basic experimental studies found that intestinal C. tropicalis promotes colorectal carcinogenesis and tumor progression by inhibiting antitumor immunity.
27
,
28
,
29
Interestingly, these studies also revealed that C. tropicalis promotes chemotherapeutic drug resistance in CRC, and antifungal drugs inhibited carcinogenesis and tumor progression of CRC in animal models.
27
,
30
Based on the results of these previous studies, we hypothesized that increased C. tropicalis burden in CRC tissue specimens might serve as a critical novel prognostic biomarker for patients with CRC.
Here, we investigated the potential of a tissue‐based fungal biomarker to predict prognosis in patients with CRC. We quantified the intratumoral C. tropicalis burden using quantitative polymerase chain reaction (qPCR) assays in CRC tissues and evaluated the association between the intratumoral burden of this fungi and cancer‐related prognosis in patients with this malignancy.

MATERIALS AND METHODS

Patient cohorts
This study included 304 clinical frozen surgical tissue specimens from patients with pathological stage (pStage) I through III CRC who underwent curative surgery at the Institute of Science Tokyo Hospital, Japan, between 2015 and 2018. All patients received curative surgery for primary tumors pathologically diagnosed as pStage I through III based on the TNM Classification of Malignant Tumors, 8th edition.
31
None of the patients received neoadjuvant chemotherapy or radiotherapy before curative surgical resection. Clinical data were collected from electronic medical records, with follow‐up continuing until death or for at least 5 years after surgery.
This study was performed in accordance with the tenets of the Declaration of Helsinki and approved by the Institutional Review Board of the Institute of Science Tokyo (M2000‐831). Written informed consent for inclusion in the study was obtained from all patients.

Genomic DNA extraction and qPCR assays
All procedures were conducted aseptically in a clean, isolated laboratory area physically separated from routine clinical processing spaces to minimize the risk of contamination. First, surgical tissue specimens were placed in sterile containers under aseptic conditions immediately after surgical resection and transported to the laboratory. Each specimen was subsequently divided into small pieces using sterile surgical blades and immediately stored at –80°C in sterile tubes. Genomic DNA was extracted from the frozen surgical tissue specimens using the RNeasy Mini Kit (Qiagen, Hilden, Germany), and the genomic DNA quality was measured using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, US). The SensiFAST SYBR Lo‐Rox Kit (Bioline, London, UK) and StepOnePlus Real‐Time PCR System (Applied Biosystems, Foster City, CA, US) were then used for qPCR. The intratumoral burden of C. tropicalis was measured using the primers for amplifying the 26S ribosomal RNA (rRNA), according to previously described methods.
32
,
33
,
34
The specificity of the primers was assessed using electrophoresis and the GenBank database, and the burden of C. tropicalis was normalized to the expression levels of the HBB gene using the ΔCt method. The primers used in this study are shown in Supporting Information S1: Table S1. In each qPCR run, multiple no‐template control wells containing nuclease‐free water were included to rigorously monitor for potential contamination.

Mycobial composition and diversity analyses
The 16S rRNA sequence dataset (34 adjacent normal and 49 CRC cases) for GSE165255 was obtained from the Gene Expression Omnibus repository browser (https://www.ncbi.nlm.nih.gov/geo/browse/). This metagenomic dataset was mapped to the Homo sapiens and fungus databases using Kraken2, and the data were assigned to each taxon. KrakenTools was then used to calculate alpha diversity and genus/species fractions.
35
Thereafter, we extracted Candida species data from the entire metagenome data and obtained the fraction information. KrakenTools was then used to calculate alpha diversity.

Microsatellite instability analysis
The microsatellite instability analysis was evaluated using five mononucleotide repeat markers, which were BAT25, BAT26, D2S123, D5S346, and D17S250, by an ABI 3130 Genetic Analyzer (Applied Biosystems), as described previously.
36
,
37

Statistical analyses
Statistical analyses were performed and graphs were generated using the appropriate packages contained in EZR version 1.68,
38
a graphical user interface for R (version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria). Fisher’s exact test was used to analyze categorical variables, and a two‐sided Student’s t test was used to analyze differences between continuous values. For all statistical analyses, significance was set at a level of 5% (p < .05).
The cutoff points for the intratumoral burden of C. tropicalis were decided using receiver operating characteristic curve analysis. Survival performance was estimated using the Kaplan–Meier method and log‐rank tests. Multivariate Cox proportional hazards regression analysis was then performed to assess recurrence and disease‐specific survival using the clinicopathological factors identified as significant in the univariate analysis.

RESULTS

Mycobial metagenomic analyses revealed significant changes of fungal diversity in Candida species in colorectal cancer tissues
First, we focused on the differences in fungal composition and diversity of Candida species between adjacent normal and cancer tissues in patients with CRC. Mycobial diversity analyses using the metagenomic dataset of the Candida species (obtained from the GSE165255 16S rRNA sequence dataset [34 adjacent normal and 49 CRC cases]) demonstrated that Simpson's reciprocal index, the Shannon diversity index, and Fisher's alpha parameter were significantly higher in CRC tissues than in adjacent normal tissues (p = .03, p = .05, and p = .03, respectively; Figure 1A), suggesting significant alternations of diversity in Candida species in CRC. Furthermore, the mycobial composition analysis of Candida species revealed a significant difference in the relative abundance of each Candida species between CRC and adjacent normal tissues (Figure 1B). Among these common Candida species, C. tropicalis demonstrated significantly higher abundance in cancer tissues than in adjacent normal tissues (p = .01; Figure 1C). Overall, the mycobial metagenomic analyses demonstrated the significant difference in fungal composition and diversity in Candida species—especially in C. tropicalis—within CRC tissues, suggesting its potential as a diagnostic and prognostic biomarker for patients with CRC.

Increased intratumoral burden of C. tropicalis was related to cancer recurrence and disease‐specific survival in CRC
To examine the prognostic biomarker potential of C. tropicalis, the intratumoral C. tropicalis burden was evaluated in surgical cancer tissues from 304 patients with pStage I–III CRC using qPCR assays. The burden was measured using a pair of specific primers for the 26S rRNA of C. tropicalis, and the patients were categorized into high and low C. tropicalis burden groups by the cutoff thresholds, which were determined using receiver operating characteristic curve analysis and Youden’s index.
39
The high and low intratumoral burden groups were identified to include 122 (40.1%) and 182 (59.9%) patients, respectively. The clinicopathological characteristics according to the intratumoral C. tropicalis burden are demonstrated in Table 1. Interestingly, the intratumoral C. tropicalis burden was not correlated with any key clinicopathological features, including patient demographics, cancer characteristics, and serum cancer markers.
Regarding recurrence‐free survival (RFS) and disease‐specific survival (DSS) in this clinical cohort, 48/304 patients (15.8%) experienced cancer recurrence during the follow‐up period and 16/304 patients (5.3%) experienced disease‐specific death. Notably, patients with CRC who experienced recurrence tended to have a higher intratumoral C. tropicalis burden when compared with those without recurrence (p = .13; Figure 2A). When stratified by major recurrence patterns, the intratumoral C. tropicalis burden was marginally higher in patients who developed peritoneal dissemination recurrence than in those without recurrence, representing the most pronounced difference among all major recurrence patterns examined (p = .05; Supporting Information S1: Figure S1). Similarly, patients who experienced disease‐specific death had a significantly higher C. tropicalis burden than those who did not experience disease‐specific death (p = .04; Figure 2B). Moreover, Kaplan–Meier analysis revealed that the group with a high intratumoral C. tropicalis burden had a significantly poorer RFS than the low burden group (5‐year RFS rate, 78.0% vs. 86.6%; p = .03; Figure 2C). In line with RFS, high burden group of intratumoral C. tropicalis exhibited a significantly worse DSS than the low burden group (5‐year DSS rate, 88.9% vs. 98.0%; p < .01; Figure 2D). A similar tendency was also observed for overall survival (Supporting Information S1: Figure S2).
Multivariate Cox regression analysis of RFS and DSS—including clinicopathological factors identified as significant in the univariate analysis—revealed that the intratumoral C. tropicalis burden was a significant and independent predictor for RFS (hazard ratio [HR]: 1.92; 95% CI, 1.08–3.44; p = .03; Table 2). Similarly, an increased intratumoral C. tropicalis burden was a significant independent predictor for DSS in patients with CRC (HR: 4.29; 95% CI: 1.36–13.5; p = .03; Table 3). Overall, increased intratumoral burden of C. tropicalis was significantly related to cancer recurrence and disease‐specific survival in CRC, and these results highlight intratumoral C. tropicalis burden have the potential to serve as novel prognostic biomarkers in patients with pStage I–III CRC.

The effect of C. tropicalis burden on RFS and DSS was not modified by any other key clinicopathological features
Subgroup analysis revealed that the effect of the intratumoral C. tropicalis burden on RFS and DSS was not significantly modified by any of the key clinicopathological features identified to be significant in the univariate analysis—age, sex, location of the cancer, tumor size, and lymph node metastasis (Figure 3A and B). However, a relative interaction was observed for pathological T stage: in patients with T3–4 disease, the HRs were 1.00 for RFS and 1.03 for DSS, whereas in those with T1–2 disease, the corresponding HRs were markedly higher at 2.15 for RFS and 2.40 for DSS, suggesting that the prognostic impact of intratumoral C. tropicalis burden may be more pronounced in patients with early‐stage disease (Figure 3A and B). Taken together, we, for the first time, demonstrated that intratumoral C. tropicalis burden are associated with tumor recurrence and cancer‐related survival, and a novel significant predictor of prognosis in patients with CRC.

DISCUSSION
In this study, mycobial metagenomic analyses revealed significant alterations in fungal composition and diversity within CRC tissues. Moreover, we found that an increased intratumoral burden of C. tropicalis was a significant and independent predictor of cancer recurrence and DSS in CRC. These results indicate that intratumoral C. tropicalis burden has the potential to serve as novel prognostic biomarker in patients with CRC.
The recent dramatic advancement of comprehensive molecular analytical techniques, including 16S rRNA sequencing, has allowed for the identification of dysbiosis caused by specific bacterial species in various diseases, including cancers. This has enabled the analytical study of microbiomes in cancer to be conducted worldwide, and a large number of studies have attempted to develop novel microbiome‐based diagnostic and prognostic biomarkers for various types of cancers, including CRC. For example, B. fragilis, E. coli, and F. nucleatum have extensively been investigated as potential biomarkers for CRC.
40
,
41
,
42
,
43
,
44
,
45
Furthermore, more recent studies have employed metagenomic sequence‐based analyses of CRC surgical tissues and fecal specimens in an effort to identify multiple novel bacterial species as biomarkers for CRC and develop microbiome‐based biomarker models for the early detection, prognosis prediction, and prediction of therapeutic efficacy in patients with CRC.
46
,
47
,
48
,
49
However, while the development of microbiome‐based biomarkers for CRC is gaining attention and progressing, and several studies have attempted to establish mycobiome‐based biomarkers for CRC,
20
,
50
the development of biomarkers targeting the fungal species remains largely unexplored. In this study, by examining the intratumoral C. tropicalis burden in clinical surgical tissue specimens, for the first time, we have identified a novel mycobiome‐based prognostic biomarker for predicting cancer‐related outcomes in patients with CRC. Unlike most previous metagenomic sequence‐based studies, our biomarker was based on qPCR technique, which is a simple, easy‐to‐use, and cost‐effective approach, enabling easy application in clinical practice.
Recently, increasing evidence that gut fungi dysbiosis plays a significant role in the carcinogenesis and progression of CRC has been accumulating, and the crucial role of C. tropicalis in the carcinogenesis and progression of CRC has attracted attention. For example, recent metagenomics‐based studies have revealed that human CRC surgical tissues and stools from patients with CRC exhibit dysbiosis of C. tropicalis.
20
,
27
Furthermore, several basic experimental studies using animal models have revealed that intestinal C. tropicalis promotes colorectal carcinogenesis and tumor progression by inhibiting antitumor immunity and regulating the tumor microenvironment.
27
,
28
,
29
,
30
Interestingly, these studies also demonstrated that chemotherapeutic drug resistance was promoted by C. tropicalis and tumor progression was inhibited by antifungal treatment against C. tropicalis. Although recent studies have attempted to clarify the molecular mechanisms underlying the association between C. tropicalis and CRC, few studies have used clinical samples to investigate the association between C. tropicalis and CRC or develop mycobiome‐based biomarkers for CRC. In this study, unlike these previous studies, using a substantial number of clinical surgical tissue specimens, we first investigated a significant association between C. tropicalis and cancer‐related survival outcomes in patients with CRC. Thus, we developed a novel mycobiome‐based prognostic biomarker for CRC.
As noted, although the difference did not reach statistical significance (p = .13; Table 1), patients with a high intratumoral C. tropicalis burden exhibited a numerically greater proportion of right‐sided tumors. Although this finding should be interpreted cautiously, it may have biological significance. Right‐ and left‐sided CRCs are well recognized to differ in tumor biology and tumor‐associated microbial ecosystems. Spatial heterogeneity of the microbiome along the colon has been documented, and microbial communities have been shown to vary according to tumor location.
23
,
51
Furthermore, fungi have been reported to exhibit regional enrichment within the colon, particularly in distal segments.
52
Taken together, these findings suggest that regional differences in the colonic microenvironment—including bile acid composition, oxygen gradients, and immune contexture—may influence fungal colonization or expansion within tumors. Nevertheless, because the observed trend did not achieve statistical significance in our cohort, this finding should be interpreted with caution and warrants further investigation.
In the subgroup analysis, the prognostic impact of high C. tropicalis burden appeared more pronounced in patients with T1–2 tumors than in those with T3–4 tumors. However, formal testing for interaction between T stage and C. tropicalis burden did not reach statistical significance; therefore, these findings should be interpreted with caution. Notably, T3–4 stage itself was a strong and independent predictor of both RFS and DSS in our cohort, suggesting that in advanced‐stage disease, tumor‐intrinsic factors may dominate overall prognosis and attenuate the relative contribution of fungal burden to survival outcomes. In contrast, in earlier stage tumors—where tumor‐driven hazards are comparatively less overwhelming—microenvironmental and host immune factors may exert proportionally greater influence. This interpretation is supported by prior evidence demonstrating that immune contexture independently modifies CRC outcomes beyond conventional TNM staging,
53
and that the immune microenvironment serves as a key determinant of cancer prognosis across tumor stages.
54
Collectively, these findings suggest that the prognostic relevance of intratumoral C. tropicalis burden may be influenced by an interplay between tumor stage–dependent risk and host immune context, highlighting tumor stage as a contextual modifier when interpreting mycobiome‐associated outcomes.
Despite its strengths, this study had some potential limitations. First, this was a retrospective study conducted at a single institution. Because of differences in diet, environmental factors, and genetic background, the composition of gut mycobiome is likely to vary across countries and regions. Therefore, to ensure the generalizability of our study findings, our results need to be validated with specimens from multiple countries and regions. Second, we did not perform internal validation by splitting the cohort into training and validation sets. Considering the total sample size of 304 patients, such division would have substantially reduced statistical power, particularly for subgroup analyses comparing patients with high and low intratumoral C. tropicalis burden. We analyzed the entire cohort as a single dataset to ensure consistency of the findings. To our knowledge, this is the first clinical study to systematically evaluate the association between intratumoral C. tropicalis burden and CRC prognosis. Accordingly, our findings should be regarded as exploratory. Independent validation in external cohorts is essential before C. tropicalis can be considered for clinical application as a prognostic biomarker. Despite these limitations, this study represents an important first step toward understanding the clinical significance of intratumoral fungal communities in CRC prognosis.

CONCLUSIONS
Through qPCR‐based analysis using clinical cancer tissue specimens, for the first time, we have successfully demonstrated that an increased intratumoral C. tropicalis burden was significantly associated with cancer recurrence and disease‐specific survival in patients with CRC. Our findings highlight the potential of the intratumoral C. tropicalis burden as a novel prognostic biomarker for patients with CRC and shed new light on the mycobiome’s significant biological role in CRC.

AUTHOR CONTRIBUTIONS

Sakiko Oba: Conceptualization; methodology; software; data curation; investigation; validation; formal analysis; visualization; resources; writing—original draft. Keisuke Okuno: Conceptualization; methodology; data curation; investigation; supervision; formal analysis; funding acquisition; visualization; project administration; writing—review and editing. Shuichi Watanabe: Methodology; software; formal analysis; data curation; writing—review and editing. Yudai Yamamoto: Resources; writing—review and editing. Ayumi Takaoka: Resources; writing—review and editing. Marie Hanaoka: Resources and writing—review and editing. Shinichi Yamauchi: Resources; writing—review and editing. Hiroyasu Kagawa: Methodology; resources; writing—review and editing; conceptualization. Masanori Tokunaga: Conceptualization; project administration; writing—review and editing; Supervision. Daisuke Ban: Project administration; writing—review and editing; supervision. Yusuke Kinugasa: Conceptualization; writing—review and editing; project administration; supervision.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.

PATIENT CONSENT STATEMENT
A written informed consent was obtained from all patients who enrolled at the present study.

PERMISSION TO REPRODUCE MATERIAL FROM OTHER SOURCES
Not applicable.

Supporting information
Supporting Information S1