Proctitis

RNA was isolated from rectal biopsies obtained during diagnostic colonoscopy using the Qiagen AllPrep RNA/DNA Mini Kit. PolyA-RNA selection, fragmentation, cDNA synthesis, adaptor ligation, TruSeq RNA sample library preparation (Illumina, San Diego, CA), and paired-end 75 bp sequencing was performed^{9 (link)}. An additional validation of the baseline rectal gene expression at diagnosis utilized the independent RISK cohort of treatment-naïve pediatric patients (55 non-IBD controls, 43 UC patients, and 92 CD patients with rectal inflammation) and single-end 75 bp mRNA sequencing^{9 (link)}. Reads were quantified by kallisto^{51 (link)}, using Gencode v24 as the reference genome and transcripts per million (TPM) as an output. We included 14,085 protein-coding mRNA genes with TPM above 1 in 20% of the samples in our downstream analysis. Only samples for which the gene expression (Y encoded genes and XIST)-determined gender matched the clinical-reported gender were included in the analyses (we excluded only one sample with unmatched gender). Four other PROTECT samples were excluded due to poor read quality. A total of 226 RNAseq samples with mean read depth of ~47 M (14 M std. deviation) were stratified into specific clinical subgroups including Ctl (n = 20), and UC (n = 206), and were substratified based on disease severity, and on histologic findings. Differentially expressed genes were determined in GeneSpring® software with fold change differences (FC) ≥ 1.5 and using the Benjamini–Hochberg false discovery rate correction (FDR, 0.001) for all analyses except for the corticosteroid response genes that was calculated out of the 712 severity genes with FDR < 0.05. Unsupervised hierarchical clustering using Euclidean distance metric and Ward’s linkage rule was used to test for groups of rectal biopsies with similar patterns of gene expression. ToppGene^{6 (link)} and ToppCluster^{7 (link)} software were used to test for functional annotation enrichment analyses of immune cell types, pathways, phenotype, immune cell-type enrichments, and biologic functions. Visualization of the network was obtained using Cytoscape.v3.0.2^{52 (link)}.
For validation of the association between baseline gene expression and outcome, we also generated independent Lexogen QuantSeq 3′ mRNAseq libraries^{19 (link)} and single-end 100 bp sequencing for 134 participants who also had Illumina mRNAseq data (the Discovery Cohort) and for 50 participants who did not have Illumina mRNAseq data (the independent Validation cohort, Table 1). PCA was performed to summarize variation in gene expression between patients, and principal components (PC) values were extracted for downstream analyses. We considered several central gene expression pathways PC1 preidentified by the previous differential expression analyses and functional annotation enrichment analyses of the core 5296 UC genes, the 712 severity genes, and the 115 corticosteroid response gene signature for the model building and associations with the microbial composition as described below. PROTECT (GSE109142) and RISK (GSE117993) rectal mRNAseq data sets were deposited into GEO.

Clinical outcomes were collected for all patients who provided a viable sample, by requesting copies of examination reports from participating sites every month. All diagnoses were determined by reviewing endoscopy, radiology, and histology reports, clinic letters, and urgent referral forms provided by the participating sites. Patient and clinical data included symptoms, reasons for the referral, medical history, and sociodemographic factors. All diagnoses were verified by medical members of the central research team.
All neoplastic bowel polyps, either adenomatous polyps or sessile serrated polyps, were identified and were given a risk of either ‘low’, ‘intermediate’, or ‘high’ depending on their size and frequency; contemporary UK and European guidelines were used in this study^{23 ,24 (link)}, with low risk defined as 1–2 adenomas less than 10 mm, intermediate risk as 3–4 small adenomas less than 10 mm or one adenoma 10 mm or more, and high risk as five or more adenomas less than 10 mm or three or more adenomas with at least one 10 mm or more.
Non-neoplastic polyps, such as hyperplastic, inflammatory, or pseudopolyps, were classified separately. Patients with a risk score for their polyps at first, second, or third examinations had their cumulative number and/or highest risk polyp taken as their final score. Remaining bowel pathology was classified as one of CRC, inflammatory bowel disease (colitis/proctitis), diverticulosis, haemorrhoids, normal examination, or procedure stopped/incomplete. Patients with concurrent polyps and CRC were classified as CRC and not included in the analysis, as our target group for this study was those without CRC in whom we could potentially identify polyps and plan for removal before they could progress to CRC.

A variety of nCRT-related adverse reactions were evaluated, including bone marrow suppression, radioactive proctitis, intestinal obstruction or perforation, narrow lumen, anastomotic fistula, perianal skin injury, emesis, and hand-foot syndrome. Hand-foot syndrome was mainly associated with capecitabine treatment. During concurrent chemoradiotherapy, blood routine examinations and biochemical examinations were conducted weekly. RTOG radiation injury classification and Common Terminology Criteria for Adverse Events (CTCAE, Version 5) were adopted to assess adverse events. Grades 1 and 2 myelosuppression were considered mild, while grades 3 and 4 were considered moderate to severe. Similarly, grades 1 and 2 were defined as mild radiation proctitis, and grades 3 and 4 were defined as moderate to severe radiation proctitis. The remaining adverse reactions including intestinal obstruction or perforation, narrow lumen, anastomotic fistula, perianal skin injury, emesis, and hand-foot syndrome were evaluated by their occurrence or not.