**3. Discussion**

To date, CC and RC are regarded as a single disease entity, termed CRC. Biologically, CRC is a heterogeneous group of tumors, characterized by high interpatient and intratumor heterogeneity with variable clinical features and outcomes. Tumor sidedness (CC versus RC) is one aspect of heterogeneity, and it correlates with distinct biological and molecular characteristics, as well as with different disease management strategies. In surgical clinical oncology, for decades researchers have been becoming aware of the fact that CC and RC are different diseases, based on their surgical procedures and challenges, as well as complication rates and local recurrence patterns.

We believe that understanding CRC heterogeneity and regarding CC and RC as two different tumor entities is fundamental to overcoming the inconsistent study results and the heterogeneous nature of microbiomes. Thus, our prospective, observational study aimed to characterize the microbiome landscape of different body sites in patients with treatment-naive CC. Moreover, we correlated the microbiome with sidedness (RSCC versus LSCC) and other clinicopathologic features of tumor progression (such as stage, lymph node involvement and tumor grade). Right or extended right hemicolectomy with complete mesocolic excision involves the resection of the tumor along with nonmalignant tissues, including the terminal ileum. These procedures provide surgical access to the ileal lumen. Studies investigating the bacterial composition of CC via a comparison of matched samples from multiple locations in the body, such as feces, tumor tissue and normal-healthy mucosa tissue, are rare and have reported inconsistent results. As mentioned before, one reason might be that the majority of these studies combined CC and RC samples as CRC for their analyses. Furthermore, analyzing the gut microbiome using stools does not capture all the microbes in the gut, in particular mucosally adherent microbes and microbes in the small intestine (ileal microbiota). Most of our knowledge has been derived from studies of ileal biopsies during colonoscopies or naso-ileal catheters. However, data must be interpreted cautiously because accessing the ileal microbiome via retrograde examinations is prone to contamination [34,35,40,42–45]. The ileal microbiota is oral-like and more variable than its colonic counterpart, and across several studies, the ileal core microbiome is constituted by *Firmicutes*, *Proteobacteria* and *Actinobacteria*. Villmones et al. reported the ileal microbiome of 27 patients based on samples collected during radical cystectomies with urinary diversion. They demonstrated that the distal part of the ileum harbors a distinct niche that differs from the colonic flora. The REIMAGINE study revealed that the stool microbiome was a good proxy for that of the large intestine but differed substantially from that of the small intestine [19]. The Zitvogel and Roberti group recently reported that the ileal immune tonus was affected by colonic carcinogenesis in RSCC, indicated by the fact that the growth of heterotopic or orthotopic CCs induced this upregulation of ileal immune gene products [34,35]. They also demonstrated that the ileal microbiome governed the efficacy of chemotherapy and PD-1 blockade in CC independent of microsatellite instability. Our study further reveals that the presence of a colonic tumor leads to a more consistent cancer-defined microbiome and shapes the normal spatial heterogeneity existing along the intestinal tract. No significant differences in alpha or beta diversity were identified between the ileal samples and tumor samples. In our cohort, due to operational reasons, in several cases of LSCC, an extended operation was needed, and from those patients, we also collected terminal ileum specimens. Interestingly, also in this subgroup, we did not observe a significant difference in beta diversity between the tumor and ileum. *Bifidobacterium* was significantly associated with LSCC and was found in the core microbiome of more than half of the ileal samples of the LSCC patients. In contrast, *Bifidobacterium* was not found to be a core microbiota in the ileal samples of the RSCC patients. Furthermore, the abundance profile of the terminal ileum revealed that samples from the patients with LSCC harbored 37% *Streptococcus* and 19% *Enterobacter*, while samples from the patients with RSCC harbored 10% *Streptococcus* and only 2% *Enterobacter*. The subgroup of the MSI patients harbored 20% *Akkermansia* and 25% *Streptococcus*, while in the MSS patients, the percentage of *Akkermansia* was less than 1%, and the amount of *Streptococcus* was 8%. A limitation of the study is that the microbiome shifts might have been induced by the preoperative bowel preparation, but colonoscopy sampling also requires bowel preparation.

However, we know that the fecal microbiota differs from the microbiota of mucosal tissue in regards to oxygen and nutrition needs [46]. Analyzing the bacterial composition, especially the similarity or dissimilarity, between tumors, healthy mucosa and stool from the same individual provides information regarding changes in the microenvironment that have occurred that favor growth in the right- or left-sided colon. Most microbiota identified from human feces belong to the phyla *Firmicutes, Bacteroidetes, Proteobacteria, Actinobacteria* and *Verrucomicrobia*, with 90% belonging to *Firmicutes* or *Bacteroidetes*. Although a diseasespecific microbiota signature has yet to be identified, patients with CRC have reduced bacterial diversity and richness compared to healthy individuals. Specific bacteria, such as Fusobacterium nucleatum, as well as certain *Bacteroides fragilis* and *E. coli* species, are known CRC-associated pathobionts. We confirm previous observations that CC tumors harbor orally derived opportunistic pathogens [47,48]. Furthermore, we observed that the stool microbiome only partially reflects the tumor microbiome. We identified 35 genera whose abundance significantly differed between tumor and stool samples. Interestingly, between paired tumor and nontumor healthy colon tissue, no significant differences were observed. We think the presence of a colonic tumor leads to a more consistent microbiota profile. This finding supports previous studies by Murphy et al. and Liu et al., which demonstrated that the microbiotas in tumor tissue and normal mucosa tissue of patients with CRC were similar [49,50]. We further observed that RSCC and LSCC patients harbor distinct microbiomes, characterized by differences in microbial diversity and bacterial taxa. The alpha diversity in the LSCC patients was significantly higher than that in the RSCC patients. Consistent with our results, Phipps et al. showed that patients with RSCC showed fewer taxonomic differences than those with left-sided carcinomas [51]. However, unlike our study, the study of Phipps et al. included rectal cancer patients. Furthermore, we were able to show that the tumor tissue of RSCC patients was characterized by a significant increase in the abundances of *Haemophilus* and *Veilonella*, while increased abundances of *Bifidobacterium* and *Ruminococcus* were associated with LSCC. Overall, grade 3 tumors were significantly enriched in *Fusobacterium* and *Parvimonas*. Little is known about *Parvimonas,* but interestingly, *Parvimonas micra* and *Fusobacterium* have been shown to aggregate and form biofilms in vitro [52,53]. Biofilm formation is linked to inflammatory bowel disease and CC. Due to dysbiosis, biofilm formation occurs within the inner mucus layer, normally free from microorganisms, which could result in direct contact between bacteria and epithelial cells [54]. As mentioned in the introduction, CRC numbers are rising in younger people worldwide. The increased incidence of early-onset CRC can be the consequence of environmental influences (e.g., having a Western diet, food quality and additive-laden food). Early onset is more frequent in left-sided colon. We consider the microbiome of someone developing colorectal cancer at an age over 80 years to be different from someone with early-onset colorectal cancer. Unfortunately, we have too few patients in the "younger age" group for a detailed analysis. In line with the group of P. O'Toole, we recommend adjusting for age to improve the identification of gut microbiome alterations in multiple diseases.

Accumulating evidence suggests a critical role of intestinal dysbiosis in surgical site infections and anastomotic leakage after CRC surgeries. Despite improvements in surgical techniques, new energy devices and intensive care management, anastomotic leakage is still a significant problem in daily clinical practice. We recently linked the microbiome to surgical complications in pancreatic surgeries [18]. In CRC surgeries, the microbiome has also been linked to postoperative complications [55–57]. Many factors beyond geography, diet and lifestyle affect tumors, independent of the microbiome composition, prior gastrointestinal surgery, antibiotic treatment or preoperative bowel preparation regimen. To prevent this type of possible bias, we designed a study in which all patients received the same preoperative bowel preparation regimen on the day prior to their surgery and perioperative antibiotic prophylaxis (most of the studies did not even report this treatment). Furthermore, we excluded upfront confounding variables, such as antibiotic usage, four weeks prior to

surgery and systemic conditions related to bowel dysfunction, and only patients who lived in Franconia (Germany) for at least six months were included. We observed that the preand postoperative stool microbiomes differed strikingly. Bacterial richness and evenness were significantly lower in the postoperative stool samples. Furthermore, postoperative stool samples were highly dysbiotic and characterized by a significant increase in the abundance of *Enterococcus*, a potentially pathogenic bacterium. These findings suggest that bowel preparation, perioperative antibiotic treatment and surgery had a major effect on the stool microbiome. We aim to begin a new study analyzing the impact of mechanical bowel preparation on the intestinal microbiome in the context of surgery and outcomes.
