Clinical stage III CRC patients who attended the National Cancer Center, Korea, between May 2017 and June 2017, were recruited for a prospective pilot study. We excluded the patients with (1) distant metastasis, (2) other bowel disorders such as inflammatory bowel disease, ischemic colitis, and tuberculous colitis, (3) a previous history of bowel resection, and (4) patients on medications that affect gut microbiology. Initially, 12 patients with clinical stage III CRC were recruited. Among these 12, one withdrew his research consent and was excluded. Thus, the remaining 11 patients were enrolled. We collected stool samples from these 11 patients before curative surgery and 6 months after, and analyzed changes in the gut microbial community of each patient. All study procedures were performed in accordance with the Declaration of Helsinki. This study was approved by the Institutional Review Board of the National Cancer Center (NCC2017-0126). Written informed consent was obtained from all patients prior to enrolment.

The first set of stool samples was collected prior to the start of bowel preparation for surgery and the second set was collected from the outpatient department 6 months postoperatively. In case the patient had to undergo adjuvant chemotherapy following surgery, stool samples were collected 1 month after chemotherapy.

Fresh stool samples (~30 mL) were collected in OMNIGene GUT tube (DNA Genotek, Ontario, Canada) and brought to the research center within 24 hours. Samples were stored at −70℃ in the laboratory prior to DNA extraction. DNA was extracted using a QIAamp DNA Stool Kit according to the manufacturer's protocol (Qiagen, Valencia, CA, USA). A nanodrop 2000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA) was used to check the quality and concentration of the DNA.

A 16S rRNA sequencing library was constructed according to the 16S metagenomics sequencing library preparation protocol (Illumina, San Diego, CA, USA) targeting the V3 and V4 hypervariable regions of the 16S rRNA gene. KAPA HiFi HotStart ReadyMix (KAPA Biosystems, Wilmington, MA, USA) and Agencourt AMPure XP system (Beckman Coulter Genomics, Brea, CA, USA) were used for PCR and purification of PCR products, respectively. The initial PCR was performed with 20 ng template DNA using region-specific primers shown to be compatible with Illumina index and sequencing adapters (forward primer: 5′-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3′; reverse primer:5′-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGT ATCTAATCC-3′). Following magnetic bead-based purification of polymerase chain reaction (PCR) products, the second PCR was performed using primers from a Nextera XT Index Kit (Illumina Inc., San Diego, CA, USA) with a limited cycle. Subsequently, purified PCR products were visualized using gel electrophoresis and quantified with a Qubit dsDNA HS Assay Kit Invitrogen (Carlsbad, CA, USA) on a Qubit 3.0 fluorometer. Pooled samples were run on an Agilent 2100 bioanalyzer Agilent Technologies (Santa Clara, CA, USA) for quality analysis prior to sequencing. Libraries were quantified via qPCR using CFX96 Real Time System Bio-Rad Laboratories (Hercules, CA, USA). After normalization, sequencing of the prepared library was conducted on the Miseq system (Illumina) with 300-bp pairedend reads.

The reads were sorted using unique barcodes for each PCR product. The barcode, linker, and primer sequences were then removed from the original sequencing reads To merge pairedend reads that were processed in a previous step, FLASH v 1.2.11 software was applied [18]. Merged reads that contained 2 or more ambiguous nucleotides, exhibited a low-quality score (average score < 20), or were shorter than 300 bp, were filtered out. Potential chimeric sequences were detected using the Bellerophon method.

Preprocessed reads from each sample were used to calculate the number of operational taxonomic units (OTUs). The number of OTUs was determined by clustering sequences from each sample via a 97% sequence identity cutoff using QIIME software v1.8.0 (University of Colorado, Boulder, CO, USA) [19]. Taxonomic abundance was evaluated with a RDP Classifier v1.1 (Ribosomal Database Project, Michigan State University Board of Trustees, East Lansing, MI, USA) using a confidence threshold of 0.8 derived from preprocessed reads for each sample. Microbial composition was normalized using the value calculated from the taxonomy abundance count divided by the number of preprocessed reads for each sample.

To measure the alpha diversity of each sample, the OTUs were analyzed using the Shannon index as follows: $\mathrm{H}\mathrm{\prime }=-{\sum }_{i=1}^S\left(p_{i}ln\left(p_i\right)\right)$. In order to measure beta diversity, the difference in organism composition was measured according to Bray-Curtis distance as follows: ${BC}_{ij}=\frac{S_i+S_j+{2C}_{ij}}{S_i+S_j}$. Principal component analysis (PCA) was then performed using the measured beta diversities. Continuous variables (OTUs and alpha diversity) were analyzed using the Student t-test or Wilcoxon rank-sum test following normality tests. The edge R software version 3.6.8 (Bioconductor, Garvan Institute of Medical Research, Darlinghurst, NSW, Australia) was used to compare the microbial composition between preoperative and postoperative stool samples. An adjusted P-value < 0.05 was considered statistically significant and all statistical analyses were performed using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA).

The demographic and clinical characteristics of the participants are shown (Table 1). Mean age was 64.3 ± 14.0 years and 7 patients (63.6%) were female. Tumors were evenly located; as 5 patients exhibited right side (cecum to transverse colon) tumors and 6 patients exhibited left side (splenic flexure to rectum) tumors. All surgeries were performed via a laparoscopic procedure. Right hemicolectomy and anterior resection were performed according to the location of the lesion. Nine patients underwent postoperative chemotherapy. Among these, 7 received FOLFOX and 2 received Capecitabine. FOLFOX (1 cycle; oxaliplatin, 85 mg/m², LV 200 mg/m² bolus 5-FU 400 mg/m² on day 1, infusional 5-FU 1200 mg/m² on day 1 and day 2) and Capecitabine (1 cycle; 1,000 mg/m², oral, twice daily, from day 1 to day 14) treatments were followed by the standardized regimen. Twelve cycles of FOLFOX with 2-week intervals in between each cycle and 8 cycles of Capecitabine with 3-week intervals in between, were conducted. Median duration of chemotherapy was 162 days (147–173 days). All patients completed chemotherapy until the last cycle, but 3 received reduced FOLFOX chemotherapy due to chemotherapy toxicity.

In all patients, the number of OTUs decreased following surgery. The median number of OTUs in postoperative stool samples was significantly smaller (2,981; range, 2,195–10,002) than that (7,920; range, 3,201–13,025) in the preoperative stool samples (P < 0.001) (Fig. 1A). In most patients (9 of 11), alpha diversity (Shannon Index) decreased after surgery. We found that mean alpha diversity of postoperative stool samples was also significantly lower (3.68 ± 0.37) than that (4.16 ± 0.51) in preoperative stool samples (P = 0.019) (Fig. 1B). However, no differences in postoperative alpha diversity were found between right hemicolectomy (3.48 ± 0.31) and anterior resection (3.86 ± 0.34) (P = 0.085) (Supplementary Fig. 1).

The PCA plot for beta diversity showed a differential pattern between preoperative and postoperative stool samples (Fig. 2).

The list of taxa showing significant differences in abundance between preoperative and postoperative gut microbiota at the genus and phylum level, respectively are shown (Tables 2, 3). At the genus level, the abundances of 15 taxa were decreased while those of 10 taxa were increased. The proportions of previously known pathogens, such as Campylobacter (preoperative vs. postoperative 0.0168% vs. 0.0003%; adjusted P = 0.006), Fusobacterium (0.9557% vs. 0.0080%; adjusted P = 0.013), Haemophilus (0.4813% vs. 0.0533%; adjusted P = 0.049), Porphyromonas (0.4802% vs. 0.0002%; adjusted P = 0.006), and Prevotella (21.9602% vs. 1.1781%; adjusted P = 0.049) were significantly decreased following surgery. At the phylum level, the proportion of Fusobacteria (preoperative vs. postoperative 0.0103% vs. <0.0001%; adjusted P < 0.001) was significantly decreased following surgery. The changes in microbial composition of preoperative and postoperative stool samples at the genus level are shown (Figs. 3 and 4). These showed that the pathogens including Prevotella and Fusobacterium were significantly reduced after surgery (Fig. 4).