Document resource
Introduction Inflammatory bowel disease (IBD) confers an increased lifetime risk of colorectal cancer (CRC), and patients are enrolled into bowel surveillance programs which aim to identify and remove cancer precursors (i.e. dysplasia). Endoscopic resection of dysplastic lesions has become central to the management of dysplasia in IBD, providing an organ-preserving alternative to colectomy. Risk stratification traditionally relies on histopathology, with en bloc and complete (R0) resection conferring a more favourable risk profile. However, it is well-established that histologically nondysplastic IBD mucosa can harbour potentially carcinogenic genomic alterations (‘field cancerisation’), therefore we hypothesise that resection margins that appearing ‘histologically clear’ may not be ‘genomically clear’ of cells with precancerous alterations.Methods We analysed a cohort of 59 ulcerative colitis patients, 10 of whom progressed to advanced neoplasia (AN) (high-grade dysplasia or CRC) in the same bowel segment within 3 years (‘progressors’), and 49 who did not (‘non-progressors’). From each patient, we microdissected histologically non dysplastic resection margins and the dysplastic lesions, extracted DNA from each region and performed low coverage whole genome sequencing (lcWGS). We generated genome-wide copy number alterations (CNA).Results Dysplasia in progressors (n=10) had a higher percentage genome alteration (PGA) compared with non-progressors (48.0% vs 17.6%, p =0.028). Higher max PGA strongly predicted progression to AN (HR 43.1, p <0.001), with good discriminative ability (concordance 0.79), though some non-progressors also had high CNA burden. In lesions with non-dysplastic resection margins, progressors (n=6) had higher max PGA in the margin compared with non-progressors (52.0% vs 4.2%, p =0.006). Stratifying margins into high PGA (top 25%) vs low (bottom 75%), strongly predicted progression to AN (HR 40.6, P <0.001, concordance 0.92). Phylogenetic analysis revealed progressors harbour clonal alterations extending into surrounding margins consistent with field cancerization.Conclusions Our work demonstrates that histologically inconspicuous IBD resection margins can harbour significant precancerous genomic changes, highlighting an important limitation of histological assessment alone. Genomic analysis of the index lesion identifies biologically higher risk dysplasia through the presence of copy number alterations, reflecting features of the lesion’s malignant potential. In contrast genomic profiling of resection margins can reveal residual field cancerization and provide a measure of genomic clearance at the resection site. Incorporating genomic measures of clearance into post resection risk assessment may help distinguish patients in whom successful resection is not accompanied by a low ongoing risk.