Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

New fecal bacterial signature for colorectal cancer screening reduces the fecal immunochemical test false-positive rate in a screening population

  • Marta Malagón,

    Roles Conceptualization, Data curation, Investigation, Methodology, Validation, Writing – original draft, Writing – review & editing

    Affiliations GoodGut SL, Girona, Spain, Institut d’Investigació Biomèdica de Girona (IDIBGI), Salt, Spain, Universitat de Girona, Girona, Spain

  • Sara Ramió-Pujol,

    Roles Conceptualization, Data curation, Investigation, Methodology, Supervision, Validation, Visualization, Writing – review & editing

    Affiliation GoodGut SL, Girona, Spain

  • Marta Serrano,

    Roles Conceptualization, Visualization, Writing – review & editing

    Affiliation GoodGut SL, Girona, Spain

  • Joan Amoedo,

    Roles Formal analysis, Methodology, Writing – review & editing

    Affiliations GoodGut SL, Girona, Spain, Universitat de Girona, Girona, Spain

  • Lia Oliver,

    Roles Formal analysis, Methodology, Writing – review & editing

    Affiliation GoodGut SL, Girona, Spain

  • Anna Bahí,

    Roles Data curation, Resources

    Affiliation Institut d’Investigació Biomèdica de Girona (IDIBGI), Salt, Spain

  • Josep Oriol Miquel-Cusachs,

    Roles Investigation, Resources

    Affiliation Consorci Hospitalari de Vic, Vic, Spain

  • Manel Ramirez,

    Roles Methodology

    Affiliation Laboratori Clínic Territorial de Girona (LCTG), Salt, Spain

  • Xavier Queralt-Moles,

    Roles Methodology, Resources

    Affiliation Laboratori Clínic Territorial de Girona (LCTG), Salt, Spain

  • Pau Gilabert,

    Roles Methodology, Resources

    Affiliation Hospital Universitari de Bellvitge (IDIBELL), l’Hospitalet de Llobregat, Spain

  • Joan Saló,

    Roles Investigation, Methodology, Resources

    Affiliation Consorci Hospitalari de Vic, Vic, Spain

  • Jordi Guardiola,

    Roles Investigation, Methodology, Resources

    Affiliation Hospital Universitari de Bellvitge (IDIBELL), l’Hospitalet de Llobregat, Spain

  • Virginia Piñol,

    Roles Conceptualization, Methodology, Resources

    Affiliations Universitat de Girona, Girona, Spain, Hospital Universitari de Girona Dr. Josep Trueta, Girona, Spain

  • Mariona Serra-Pagès,

    Roles Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing

    Affiliation GoodGut SL, Girona, Spain

  • Antoni Castells,

    Roles Supervision, Validation, Visualization, Writing – review & editing

    Affiliation Gastroenterology Department, Hospital Clínic, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), University of Barcelona, CIBERehd, Barcelona, Spain

  • Xavier Aldeguer,

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Supervision, Validation, Visualization, Writing – review & editing

    Affiliations GoodGut SL, Girona, Spain, Institut d’Investigació Biomèdica de Girona (IDIBGI), Salt, Spain, Hospital Universitari de Girona Dr. Josep Trueta, Girona, Spain

  •  [ ... ],
  • L. Jesús Garcia-Gil

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliations GoodGut SL, Girona, Spain, Universitat de Girona, Girona, Spain

  • [ view all ]
  • [ view less ]


Guidelines recommend routine screening for colorectal cancer (CRC) in asymptomatic adults starting at age 50. The most extensively used noninvasive test for CRC screening is the fecal immunochemical test (FIT), which has an overall sensitivity for CRC of approximately 61.0%-91.0%, which drops to 27.0%-67.0% for advanced adenomas. These figures contain a high false-positive rate and a low positive predictive value. This work aimed to develop a new, noninvasive CRC screening tool based on fecal bacterial markers capable of decreasing FIT false-positive rates in a FIT-positive population. We defined a fecal bacterial signature (RAID-CRC Screen) in a proof-of-concept with 172 FIT-positive individuals and validated the obtained results on an external cohort of 327 FIT-positive subjects. All study participants had joined the national CRC screening program. In the clinical validation of RAID-CRC Screen, a sensitivity of 83.9% and a specificity of 16.3% were obtained for the detection of advanced neoplasm lesions (advanced adenomas and/or CRC). FIT 20 μg/g produced 184 false-positive results. Using RAID-CRC Screen, this value was reduced to 154, thus reducing the false-positive rate by 16.3%. The RAID-CRC Screen test could be implemented in CRC screening programs to allow a significant reduction in the number of colonoscopies performed unnecessarily for FIT-positive participants of CRC screening programs.


Colorectal cancer (CRC) is the second leading cause of cancer death worldwide, accounting for more than 880,000 deaths in 2018 [1]. The highest incidence rates are found in the most developed countries (i.e., Australia, New Zealand, and countries in Europe and North America), while the lowest rates are found in the poorest countries (i.e., those in Africa and Southcentral Asia). These geographic differences could be associated with Western diets and other environmental aspects. Low socioeconomic status is also related to a higher risk for developing CRC [2].

CRC mortality rates have declined progressively in many Western countries as governments have begun focusing their efforts at different prevention levels. Primary prevention aims to decrease the risk of developing colorectal neoplasms by following a healthy diet and lifestyle and chemoprevention [3]. CRC screening is a secondary prevention strategy for reducing incidence and mortality by detecting and removing precursor lesions such as colorectal adenomas [4]. A tertiary prevention strategy is CRC surveillance, which seeks to minimize the impact of previously established colorectal neoplasms on the prognosis of the patient [5].

CRC is a good candidate for disease screening, as it constitutes a significant health issue, its natural history has been widely studied, diagnostic methods for its early detection are available and its treatment is more effective when it is diagnosed early [6]. Moreover, screening programs for CRC have been demonstrated to be more cost-effective than other screening programs, such as those for breast and cervix cancers [7]. Improvement in screening quality and adherence, however, is pivotal for further reducing the prevalence of CRC-associated death. The most extensively used CRC screening strategies are based on endoscopic examinations (i.e., colonoscopy and flexible sigmoidoscopy) and fecal tests (i.e., the guaiac-based fecal occult blood test and fecal immunochemical test (FIT)). Colonoscopy is currently used as a CRC screening method in countries such as Germany and Poland. Although its sensitivity and specificity for advanced neoplasia (AN; i.e., advanced adenoma (AA) and/or CRC) approaches 100% in experienced hands [8], it also has some limitations: it requires exhaustive bowel cleansing and is time-consuming, painful, and expensive [9]. To overcome these limitations, there is an increasing interest in the use of a two-step screening strategy by first using noninvasive screening and then after a positive result, performing the colonoscopy. The most commonly used noninvasive method is FIT, which specifically detects human hemoglobin in feces. The quantitative nature of FIT allows the selection of an optimal cut-off concentration for the desired target population.

Different screening programs have been implemented in regions of Spain for those aged 50-69 years. Most of these programs first use FIT (cut-off at 20 μg of hemoglobin/g of feces), followed by a colonoscopy when a positive FIT result is obtained. In the Catalonia region, in 2013, 47.8% of the invited population agreed to participate in the CRC screening program, of which 5.4% tested positive [10].

Data from 19 studies in which FIT performance in CRC screening was assessed showed that the overall sensitivity and specificity for CRC were 79.0% and 94.0%, respectively [11]. Importantly, CRC screening strategies need to show high performance in detecting not only CRC but also precancerous lesions. In this setting, FIT sensitivity for AAs has been shown to be 28.0%, while its positive predictive value (PPV) has been shown to be 13.3%, which results in a high false-positive rate for AN [12]. False-positive results lead to patient concerns and unnecessary colonoscopies, with associated potential risks and costs. Therefore, in order to reduce false-positive results, it is necessary to develop new noninvasive tools for CRC screening with higher specificity.

In recent years, growing attention has been given to the role of microbiota in carcinogenesis. Microorganisms are thought to be involved in up to 20.0% of cancers [13], specifically colorectal cancer [14]. The intestinal mucosa is constantly exposed to the gut microbiota and its derived metabolites, which have been observed to stimulate the immune response with the potential to cause inflammation [15]. Several research groups have sequenced the 16S rRNA genes of bacteria from colonic mucosa or feces and reported that patients with CRC show colorectal dysbiosis [1619]. Some of the identified species are suspected to be involved in colorectal tumorigenesis. These species include Gemella morbillorum (GMLL), Peptostreptococcus stomatis (PTST), and Bacteroides fragilis (BCTF). All these data suggest that the fecal gut microbiota is a potential alternative noninvasive tool for CRC screening. From these studies, we designed quantitative polymerase chain reaction (qPCR) systems that specifically target those bacterial markers. Moreover, we have also studied other bacterial indicators retrieved from other studies and previously developed by our group (i.e., B10, best BLAST match Faecalibacterium prausnitzii (FPRA); B46, best BLAST match Subdoligranulum variabile; and B48, best BLAST match Ruminococcus sp., Roseburia sp., and Coprococcus sp.) [20,21].

In this work, we developed a new, noninvasive tool for CRC screening (i.e., RAID-CRC Screen) based on a fecal bacterial signature that complements FIT and increases its specificity and PPV for AN detection among FIT-positive participants. Proof-of-concept and clinical validation of the designed tool in two independent cohorts of an organized, population-based CRC screening program are reported.


Study population

Participants of the Catalan CRC screening program were invited to participate in the study. The inclusion criteria were as follows: (1) asymptomatic subjects; (2) subjects aged 50–69 years (both included); and (3) subjects with a positive FIT result (cut-off at 20 μg of hemoglobin/g of feces). The exclusion criteria were as follows: (1) subjects who had received antibiotics the month prior to inclusion; (2) subjects who had received chemotherapy and/or radiotherapy the last 6 months prior to inclusion; (3) subjects with severe comorbidity that, in the opinion of the investigator, should preclude participation in the study; and (4) subjects who were pregnant at the time of inclusion.

In the proof-of-concept study, a cohort consisting of 189 consecutive FIT-positive participants in the Catalan CRC screening program was recruited (Table 1). The recruiting centers were the Hospital Universitari Dr. Josep Trueta-IAS (Girona, Spain), the Hospital Universitari de Bellvitge (L’Hospitalet de Llobregat, Spain), and the Consorci Hospitalari de Vic (Vic, Spain). The study protocol (clinical investigation code: RAID-CRC 20202015) was approved by the Clinical Research Ethics Committee of the three participating centers. Written informed consent was obtained from all participants. From 189 recruited subjects, 17 had to be excluded because of poor sample condition. Finally, 172 samples from asymptomatic subjects were used in the proof-of-concept study.

Table 1. Characteristics of patients included in the proof-of-concept (RAID-CRC 20202015) and clinical validation (GG-RAIDCRC-1002) studies, classified according to their diagnosis.

All subjects underwent colonoscopy to determine their colorectal status. According to endoscopic examination and pathology results, diagnosis was classified into four groups: normal colonoscopy (colonoscopy with no findings or with sigma and/or rectum hyperplastic polyps <10 mm), nonadvanced adenomas (NAAs; tubular adenomas <10 mm with low-grade dysplasia), AAs (adenomas ≥10 mm or with villous component or high-grade dysplasia, and pTis adenocarcinoma), and invasive CRC. Serrated lesions were classified in the NAA (lesions <10 mm without dysplasia) or AA (lesions ≥10 mm or with dysplasia) groups. Advanced neoplasms were defined as AA or invasive CRC. Individuals were also asked to answer a questionnaire to record clinical and epidemiologic data.

This proof-of-concept allowed us to develop an algorithm that discriminates between healthy subjects and those with AN. It was validated on an independent cohort consisting of 359 consecutive FIT-positive participants in the national CRC screening program (Table 1). Clinical validation was designed as a cross-sectional, multicenter study; the protocol (clinical investigation code: GG-RAIDCRC-1002) was approved by the Clinical Research Ethics Committee of Hospital Universitari Dr. Josep Trueta-IAS and Hospital Universitari de Bellvitge. Written informed consent was obtained from all study participants. According to the endoscopic findings, subjects were classified into the four groups previously mentioned. From the 359 recruited subjects, 32 had to be excluded from the study for different reasons: the subject did not collect the stool sample, the sample was incorrectly conserved, there was no colonoscopic diagnosis, and the sample condition was poor. Finally, 327 samples were used for the clinical validation of RAID-CRC Screen.

Fecal sample collection

Participants were asked to collect a stool sample from one bowel movement in a sterile container before colonoscopy and prior to bowel cleansing. Samples were immediately frozen after deposition. Then, subjects brought samples to the hospital, where they were kept frozen at -20°C for short-term storage and stored at -80°C upon arrival at the GoodGut S.L. facilities in Girona (Spain).

DNA extraction from stool samples

Genomic DNA was extracted from frozen fecal samples after homogenization using the NucleoSpin® Soil Kit (Macherey-Nagel GMbH & Co., Duren, Germany). The instructions of the manufacturer were followed, and the DNA was finally eluted in a 100 μl final volume of SE Elution Buffer and stored at -20°C until use. The DNA concentration was determined with Qubit fluorometric quantification (Thermo Fisher Scientific, Massachusetts, USA).

qPCR assay for CRC biomarkers

The specific bacterial sequences targeted were classified into five different groups according to their characteristics: Eubacteria (EUB) as the total bacterial load [22]; B10 (best BLAST match F. prausnitzii), B46 (best BLAST match S. variabile), B48 (best BLAST match Ruminococcus, Roseburia, Coprococcus), FPRA [23], and Roseburia intestinalis (RSBI) as butyrate-producing bacterial biomarkers; GMLL, PTST, and BCTF as opportunistic pathogens usually found in the oral cavity or in the bowel of mammals; Bacteroides thetaiotaomicron (BCTT) as a saccharolytic species; and Escherichia coli (ECO) as a proinflammatory species [24].

Quantification of CRC-specific biomarkers was performed by preparing specific reactions for each biomarker using SYBR Green Master Mix (Promega, Madison, USA) or Probe Master Mix (Promega, Madison, USA). The sequences of the forward and reverse primers and probes (when applied) are described in S1 Table. Each reaction consisted of 10 μl containing 1× GoTaq qPCR Master Mix, between 150 nM and 300 nM of each primer and/or probe, and up to 10 ng of genomic DNA template. The species-specific primers were purchased from Macrogen (Macrogen, Seoul, South Korea). All qPCR was run on an AriaMx Real-time PCR System (Agilent Technologies, Santa Clara, USA). Thermal profiles were different according to the biomarker analyzed (S2 Table). A melting curve step was added to the end of each qPCR to verify the presence of the expected amplicon size as well as to control primer dimer formation. Data were collected and analyzed with Aria Software version 1.5 (Agilent Technologies, Santa Clara, USA). All samples were amplified in duplicate and were considered valid when the difference between threshold cycles (Ct) was less than 0.6. A no-template control reaction and a standard curve (sequential dilutions from 107 to 103 genomic units/μL) were included in each PCR run to check qPCR efficiency.

In terms of qualitative analysis, the absence of a biomarker was considered if the obtained Ct value was not within its dynamic range, which is the interval of relative abundance in which a given bacterial marker can exist.

Statistical analysis

Data normality was assessed through the Kolmogorov-Smirnov test. The nonparametric Kruskal-Wallis test was used to test differences in variables with more than two categories. Pairwise comparisons of subcategories of these variables were analyzed using a Mann-Whitney test. All comparisons using bacterial markers were performed between the relative abundances, which were normalized by the dynamic range of each bacterial marker.

Receiver operating characteristic (ROC) curve analysis was applied to determine the usefulness of each biomarker to distinguish among different colorectal neoplasia statuses. The accuracy of discrimination of each bacterial marker was measured by the area under the ROC curve. Since RAID-CRC Screen is a qualitative approach, the performance of the designed algorithm was calculated using likelihood ratios.

In the proof-of-concept study, machine learning was used to determine which combination of bacterial markers was capable of distinguishing subjects with AN from those with normal colonoscopy or NAA. The specific methodology consisted of an initial training iteration with 70% of 100 random partitions of the dataset and further validation with 30% of the remaining predictive models generated using 4 different machine learning algorithms (neural network, logistic regression, gradient boosting tree, and random forest). RAID-CRC Screen was eventually designed using a combination of six of the bacterial markers.

The clinical validation was performed in 359 subjects. Sample size was calculated using the online platform GRANMO v7.12, estimating the population by proportions, with a 95% confidence interval, an accuracy of +/- 5%, estimating a population percentage that is expected to be approximately 30%, with 10% repositions foreseen.

The sensitivity, specificity, PPV, negative predictive value, and accuracy of the designed algorithm were calculated using Epidat 3.1 software (SERGAS, Xunta de Galicia, Spain).

All remaining statistical analyses were performed using the SPSS 23.0 statistical package (IBM, NYC, USA). Significance levels were established for p-values ≤ 0.05.


Proof-of-concept study

Fecal bacterial markers in neoplasia progression.

In the proof-of-concept study, stool samples were used to evaluate the relative abundance of each bacterial marker according to each diagnosis (Fig 1). Regardless of the colonoscopy result, the most prevalent bacterial species were butyrate producers (B10, B46, B48, FPRA, and RSBI), with relative abundance values of 16.7%, 14.2%, 16.1%, 14.1%, and 10.8%, respectively. Compared to healthy individuals, subjects with either NAA or AA showed a higher absolute abundance of B10, B46 and FPRA (p = 0.018, p = 0.004, and p = 0.023, respectively). PTST and BCTF absolute abundance was significantly higher in the CRC population than in individuals with normal colonoscopy (p = 0.002 and p = 0.017, respectively). Although there were no significant differences, a tendency could be observed in GMLL, as it showed a higher absolute abundance in CRC patients (p = 0.073); similarly, PTST and RSBI absolute abundances were higher in the presence of either NAA or AA (p = 0.056 and 0.060, respectively). Regarding ECO, no significant differences were observed among the studied groups.

Fig 1. Relative abundance of the analyzed bacterial markers in subjects with normal colonoscopy (NC), nonadvanced adenoma (NAA), advanced adenoma (AA), and colorectal cancer (CRC).

F. prausnitzii (FPRA), G. morbillorum (GMLL), P. stomatis (PTST), B. fragilis (BCTF), B. thetaiotaomicron (BCTT), R. intestinalis (RSBI) and E. coli (ECO)).

On the one hand, higher absolute abundances of opportunistic pathogens (GMLL, PTST, and BCTF) were observed in CRC patients with respect to healthy subjects (p = 0.073, p = 0.002, and p = 0.017, respectively). In terms of prevalence, these bacterial markers were found more often in patients with CRC (GMLL, 75.0%; PTST, 50.0%; and BCTF, 66.6%) than in healthy subjects (GMLL, 40.8%; PTST, 11.1%; and BCTF, 31.3%). On the other hand, when subjects with normal colonoscopy were compared to patients with AN, three butyrate-producing bacteria (B10, B46, and FPRA) showed significant differences in their absolute abundance (p = 0.035, p = 0.030, and p = 0.042, respectively).

RAID-CRC Screen algorithm development.

The development of the RAID-CRC Screen algorithm was aimed at reducing false-positive results for AN among FIT-positive subjects while maintaining 100% sensitivity for CRC. FPRA, B46, and B10 were the only bacterial markers that showed significant differences between subjects with normal colonoscopy or NAA and those with AN within the FIT-positive population; however, they did not provide the desired sensitivity values. An algorithm based on the combination of FPRA and B46 with three other bacterial species (B48, GMLL, and BCTF) and the total bacterial load exhibited a sensitivity of 94.8%, a specificity of 26.1%, a PPV of 50.0%, and a negative predictive value of 86.0% for AN (Table 2). More specifically, RAID-CRC Screen consists of 5 bacterial ratios: FPRA/EUB, B46/EUB, B48/EUB, GMLL/EUB, and BCTF/EUB applied to the FIT-positive population. While FIT 20 μg/g led to 94 false-positive results for AN detection, RAID-CRC Screen reduced this value to 62, implying a reduction of 34% in the false-positive rate. Notably, developers focused on sensitivity for AN, which reached 94.8%.

Table 2. Diagnostic performance of RAID-CRC Screen in the proof-of-concept study (n = 172).

Clinical validation study

When RAID-CRC Screen was scaled up to the 327 FIT-positive individuals of the validation cohort, sensitivities of 94.7%, 82.3%, and 83.9% were obtained for the detection of CRC, AA and AN, respectively (Table 3). Specificities were 16.3% among participants with NAA or normal colonoscopy and 18.2% among those with negative results on colonoscopy (Table 3). In this validation study, the algorithm detected 30 true negative subjects but generated 23 false negatives from the 327 FIT-positive individuals, 22 of them with AA and 1 with CRC. More importantly, while using FIT 20 μg/g, there were 184 false-positive results for AN, and RAID-CRC Screen reduced this figure to 154, which implies a reduction of 16.3% in the false-positive rate.

Table 3. Diagnostic performance of RAID-CRC Screen in the validation study (n = 327).


Early detection of CRC is crucial for reducing its incidence and mortality. The European Union recommends population-based, organized screening for CRC using evidence-based methods with quality assurance of the entire screening process [25]. The best CRC screening strategy has not yet been defined, but the most widely used strategy in Western countries is FIT, which shows 79.0% sensitivity and 99.0% negative predictive value for CRC detection [12]. The most common limitation of FIT-based CRC screening programs is false-positive results [26], as they lead to patient concerns, additional costs [27], and adverse events associated with unnecessary colonoscopies. In this study, we have defined and clinically validated a fecal bacterial signature that complements FIT by reducing its associated false-positive rate among FIT-positive participants.

Different bacterial markers associated with CRC have been analyzed in a proof-of-concept study in subjects with colorectal neoplastic lesions at different stages: normal colonoscopy, NAA, AA and CRC. Our qPCR results indicate the existence of microbial dysbiosis in patients with CRC. The analyzed bacterial markers were classified according to their gut health-related phenotypes: butyrate-producing bacteria (B10, B46, B48, FPRA, RSBI), saccharolytic bacteria (BCTT), opportunistic pathogens (GMLL, PTST, BCTF), and proinflammatory bacteria (ECO) (Fig 2). The relative abundance of these microbiological groups differed significantly when healthy subjects were compared to CRC patients. Specifically, compared to subjects with normal colonoscopy, those with CRC exhibited a decreased abundance of butyrate-producing bacteria, which were replaced by pathogenic species. Unlike other studies, ECO did not display variations across different diagnostics [28]. Although BCTF has been considered a carcinogenesis driver, it was found in higher abundances solely in CRC [29,30].

Fig 2. Relative abundance of the analyzed bacterial markers in subjects with normal colonoscopy (NC), nonadvanced adenoma (NAA), advanced adenoma (AA), and colorectal cancer (CRC).

Butyrate-producing species (B10, B46, B48, FPRA, and RSBI); opportunistic pathogens (GMLL, PTST, BCTF); saccharolytic bacteria (BCTT); proinflammatory (ECO).

The combination of these data with the FIT value was used to develop a screening method capable of discriminating between healthy subjects and patients with AN. The correct classification of subjects with CRC was prioritized; thus, a 100% sensitivity was sought. The algorithm was applied only to subjects with an FIT20 positive result (cut-off 20 μg hemoglobin/g feces) and was intended to maximize the reduction in FIT20 false-positive results while maintaining a high sensitivity for AN.

The defined microbiological signature combines five bacterial marker abundance ratios (FPRA/EUB, GMLL/EUB, B46/EUB, BCTF/EUB, B48/EUB). Our results showed that high FPRA/EUB, GMLL/EUB, and BCTF/EUB bacterial ratios correlated with the presence of AN. In contrast, low B46/EUB and B48/EUB bacterial ratios correlated with healthy individuals and AN, respectively. Notably, the use of the total bacterial load for data normalization is critical for controlling variables associated with qPCR, thus differentiating true biological changes from experimentally induced variation [31].

Clinical validation of RAID-CRC Screen confirmed high sensitivity (94.7%) and negative predictive values (96.8%) for CRC detection among FIT-positive individuals, respectively. Concerning the detection of AA, promising results were also obtained with a sensitivity of 82.3%. Altogether, RAID-CRC Screen reached a sensitivity and specificity of 83.9% and 18.2%, respectively, for the detection of AN. More importantly, the use of this new noninvasive tool achieved a reduction in unnecessary colonoscopies of up to 20% among subjects who tested positive for FIT. In addition, cancer screening programs seek methods with high negative predictive values since all affected individuals should be detected.

Although the two-step approach (FIT-positive result plus colonoscopy) is an efficient CRC screening strategy in an average-risk population [32,33], the use of bacterial signatures after testing positive in FIT may increase cost-effectiveness, as the number of unnecessary colonoscopies would be further reduced. Moreover, the use of fecal bacterial signatures might facilitate the implementation of CRC screening programs in resource-deprived regions, where colonoscopy availability is limited. The geographical origin of sample donors has been a matter of controversy that could impact the reproducibility of fecal bacterial signatures. This issue has been recently assessed by some research groups by performing meta-analyses with cohorts from different origins (USA, Germany, France, Italy, China, Canada, and Austria), and were able to identify reproducible microbiome markers in patients with CRC [19,34].

Since RAID-CRC Screen would be implemented in CRC screening organized programs, sample collection, conservation, and transport must be adapted to the existing circuits. Therefore, we have set up the detection of RAID-CRC Screen by qPCR from a unique sample taken with the FIT tube collector [20]. Thus, the availability of this screening test is linked to that of the general screening programs based on FIT. As for the cost, it will be slightly over the cost of FIT, considering further processing, but taking into account the reduction of colonoscopies it still will be more cost-effective than the current system.

Based on our results, we propose a three-step CRC screening strategy in which the bacterial signature (RAID-CRC Screen) is applied to FIT-positive individuals for a better selection of those who should undergo colonoscopy. This approach was associated with a significant reduction in false-positive FIT results among participants in a population-based, organized CRC screening program. Therefore, RAID-CRC Screen is being postulated as a new noninvasive tool for CRC screening, adding specificity and PPV to FIT while maintaining high sensitivity for AN.

Supporting information

S1 Table. Forward and reverse primers, and probe sequences used in this work.

EUB, Eubacteria; GMLL, G. morbillorum; PTST, P. stomatis; BCTF, B. fragilis; BCTT, B. thetaiotaomicron; RSBI, R. intestinalis; FPRA, F. prausnitzii; ECO, E. coli; F, Forward primer; R, Reverse primer; PR, probe. FPRA probe was 5’-labelled with FAM (6-carboxyfluorescin) as the reporter dye and TAMRA was used as quencher dye at the 3’-end.


S2 Table. qPCR conditions.

EUB, Eubacteria; GMLL, G. morbillorum; PTST, P. stomatis; BCTF, B. fragilis; BCTT, B. thetaiotaomicron; RSBI, R. intestinalis; FPRA, F. prausnitzii; ECO, E. coli; NA, nonapplicable.



We thank Ms. Alba Pla, Ms. Alba Costa, Ms. Roser Coll, Ms. Jessica Fernandez, Mr. Guillem Clos (Gastroenterology Department, Institut d’Assistència Sanitària (IAS), Salt); Ms. Dolors Serinanell and Ms. Núria Niubó (Gastroenterology Department, Consorci Hospitalari de Vic); and Ms. Elena Sanchez and Ms. Meritxell de la Hera (Gastroenterology Department, Hospital Universitari de Bellvitge) for their nursery assistance. We are grateful to Dr. Maria Buxó (Institut d’Investigació Biomèdica de Girona (IDIBGI)) for her statistical assistance. We appreciate the generosity of the patients who freely gave their time and samples to make this study possible.


  1. 1. World Health Organization. GLOBOCAN 2018. 2018. Available:
  2. 2. Doubeni CA, Laiyemo AO, Major JM, Schootman M, Lian M, Park Y, et al. Socioeconomic status and the risk of colorectal cancer: An analysis of over one-half million adults in the NIH-AARP Diet and Health Study. Cancer. 2012;118: 3636–3644. pmid:22898918
  3. 3. Castells A. Choosing the optimal method in programmatic colorectal cancer screening: current evidence and controversies. Therap Adv Gastroenterol. 2015;8: 221–233. pmid:26136839
  4. 4. Tárraga P, Solera J, Rodríguez-Montes JA. Primary and Secondary Prevention of Colorectal Cancer. Clin Med Insights Gastroenterol. 2014;7: 33–46. pmid:25093007
  5. 5. Zauber AG, Winawer SJ, O MJ, Lansdorp-Vogelaar I, van Ballegooijen M, Hankey BF, et al. Colonoscopic Polypectomy and Long-Term Prevention of Colorectal-Cancer Deaths. N Engl J Med. 2012;366: 687–696. pmid:22356322
  6. 6. Andermann A, Blancquaert I, Beauchamp S, Déry V. Revisiting Wilson and Jungner in the genomic age: a review of screening criteria over the past 40 years. Bull World Health Organ. 2008;86: 317–319. pmid:18438522
  7. 7. López Bastida J, Sassi F, Bellas Beceiro B, García Pérez L. Análisis coste-efectividad del cribado del cáncer colorrectal en la población general. Plan de Calidad para el SNS del MSC. Servicio de Evaluación del Servicio Canario de Salud. Informes de Evaluación de Tecnologías Sanitarias: SESCS N° 2006/23. 2010.
  8. 8. Groenen PJTA, Van Krieken JH, Murray S, Rouleau E, Garborg , Holme Ø, et al. Current status of screening for colorectal cancer. Ann Oncol. 2013;24: 1963–1972. pmid:23619033
  9. 9. Stracci F, Zorzi M, Grazzini G, Efird JT. Colorectal cancer screening: tests, strategies, and perspectives. Front Public Heal. 2014;2: 1–9. pmid:25386553
  10. 10. Canal Salut. Gencat. Estadístiques del programa de detecció precoç del càncer de còlon i recte. [cited 5 Aug 2019]. Available:
  11. 11. Song L-L, Li Y-M. Current noninvasive tests for colorectal cancer screening: An overview of colorectal cancer screening tests. World J Gastrointest Oncol. 2016;8: 793–800. pmid:27895817
  12. 12. Chiu HM, Lee YC, Tu CH, Chen CC, Tseng PH, Liang JT, et al. Association between early stage colon neoplasms and false-negative results from the fecal immunochemical test. Clin Gastroenterol Hepatol. 2013;11: 832–838. pmid:23376002
  13. 13. Zur Hausen H. The search for infectious causes of human cancers: Where and why. Virology. 2009;392: 1–10. pmid:19720205
  14. 14. Collins D, Hogan AM, Winter DC. Microbial and viral pathogens in colorectal cancer. Lancet Oncol. 2011;12: 504–512. pmid:21067973
  15. 15. Arthur JC, Perez-Chanona E, Mühlbauer M, Tomkovich S, Uronis JM, Fan T-J, et al. Intestinal inflammation targets cancer-inducing activity of the microbiota. Science (80-). 2012;338: 120–123. pmid:22903521
  16. 16. Sobhani I, Tap J, Roudot-Thoraval F, Roperch JP, Letulle S, Langella P, et al. Microbial Dysbiosis in Colorectal Cancer (CRC) Patients. PLoS One. 2011;6: e16393. pmid:21297998
  17. 17. Kostic AD, Gevers D, Pedamallu CS, Michaud M, Duke F, Earl AM, et al. Genomic analysis identifies association of Fusobacterium with colorectal carcinoma. Genome Res. 2012;22: 292–298. pmid:22009990
  18. 18. Zeller G, Tap J, Voigt AY, Sunagawa S, Kultima JR, Costea PI, et al. Potential of fecal microbiota for early-stage detection of colorectal cancer. Mol Syst Biol. 2014;10: 766–766. pmid:25432777
  19. 19. Thomas AM, Manghi P, Asnicar F, Pasolli E, Armanini F, Zolfo M, et al. Metagenomic analysis of colorectal cancer datasets identifies cross-cohort microbial diagnostic signatures and a link with choline degradation. Nat Med. 2019;In press. pmid:30936548
  20. 20. Malagón M, Ramió-Pujol S, Serrano M. Reduction of faecal immunochemical test false ‐ positive results using a signature based on faecal bacterial markers. Aliment Pharmacol Ther. 2019;49: 1410–1420. pmid:31025420
  21. 21. Mas de Xaxars Rivero T. Descripció i quantificació de la microbiota intestinal associada al càncer colorectal (Doctoral Thesis). 2012.
  22. 22. Matsuda K, Iwaki KK, Garcia-Gomez J, Hoffman J, Inderlied CB, Mason WH, et al. Bacterial Identification by 16S rRNA Gene PCR-Hybridization as a Supplement to Negative Culture Results. J Clin Microbiol. 2011;49: 2031–2034. pmid:21430102
  23. 23. Lopez-Siles M, Martinez-Medina M, Busquets D, Sabat-Mir M, Duncan SH, Flint HJ, et al. Mucosa-associated Faecalibacterium prausnitzii and Escherichia coli co-abundance can distinguish Irritable Bowel Syndrome and Inflammatory Bowel Disease phenotypes. Int J Med Microbiol. 2014;304: 464–475. pmid:24713205
  24. 24. Huijsdens XW, Linskens RK, Mak M, Meuwissen SGM, Vandenbroucke-Grauls CMJE, Savelkoul PHM. Quantification of Bacteria Adherent to Gastrointestinal Mucosa by Real-Time PCR. J Clin Microbiol. 2002;40: 4423–4427. pmid:12454130
  25. 25. Council of the European Union. Council Recommendation of 2 December 2003 on cancer screening (2003/878/EC). 2003.
  26. 26. Bobridge A, Bampton P, Cole S, Lewis H, Young G. The psychological impact of participating in colorectal cancer screening by faecal immuno-chemical testing–the Australian experience. Brithish J Cancer. 2014;111: 970–975. pmid:24983363
  27. 27. Lafata JE, Simpkins J, Lamerato L, Poisson L, Divine G, Johnson CC. The Economic Impact of False-Positive Cancer Screens. Cancer Epidemiol Biomarkers Prev. 2004;13: 2126–2158. pmid:15598770
  28. 28. Wassenaar TM. E. coli and colorectal cancer: a complex relationship that deserves a critical mindset. Crit Rev Microbiol. 2018;44: 619–632. pmid:29909724
  29. 29. Toprak NU, Yagci A, Gulluoglu BM, Akin ML, Demirkalem P, Celenk T, et al. A possible role of Bacteroides fragilis enterotoxin in the aetiology of colorectal cancer. Clin Microbiol Infect. 2006;12: 782–786. pmid:16842574
  30. 30. Purcell RV, Pearson J, Aitchison A, Dixon L, Frizelle FA, Keenan JI. Colonization with enterotoxigenic Bacteroides fragilis is associated with early-stage colorectal neoplasia. PLoS One. 2017;12: e0171602. pmid:28151975
  31. 31. Vandesompele J, De Preter K, Pattyn I, Poppe B, Van Roy N, De Paepe A, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002;3: 34–41. pmid:12184808
  32. 32. Wong MC, Ching JYL, Chan VCW, Sung JJ. The comparative cost-effectiveness of colorectal cancer screening using faecal immunochemical test vs. colonoscopy. Nat Sci Reports. 2015;5: 1–9. pmid:26338314
  33. 33. Quintero E, Castells A, Bujanda L, Cubiella J, Salas D, Lanas Á, et al. Colonoscopy versus Fecal Immunochemical Testing in Colorectal-Cancer Screening. N Engl J Med. 2012;366: 697–706. pmid:22356323
  34. 34. Yu J, Feng Q, Wong SH, Zhang D, Yi Liang Q, Qin Y, et al. Metagenomic analysis of faecal microbiome as a tool towards targeted non-invasive biomarkers for colorectal cancer. Gut. 2017;66: 70–78. pmid:26408641