Enhanced Neoplasia Detection in Chronic ulcerative colitis: results of the ENDCaP-C Diagnostic accuracy study

BackgroundChronic ulcerative colitis is an inflammatory condition associated with a pro-neoplastic drive, predisposing to colorectal cancer. Repeated colonoscopy is undertaken to detect preneoplastic change, but cancer diagnosis is still frequently missed. AimsTo determine if a predetermined panel of methylation markers could better risk stratify patients, aiding earlier detection of neoplasia. MethodsENDCaP-C (https://doi.org/10.1186/ISRCTN81826545) was a prospective multicentre test accuracy study of enhanced large bowel neoplasia detection and cancer prevention in patients with chronic ulcerative colitisAll patients underwent baseline colonoscopy and biopsies that had (on central review) shown no dysplasia on histology were put forward for methylation testing. In a prespecified subgroup of 200 patients without initial dysplasia detection, a second colonoscopy was performed, after 12 months. Results818 patients underwent a baseline colonoscopy. The methylation assay at baseline (testing non-neoplastic mucosa) was compared with pathology assessment at baseline for neoplasia and showed a diagnostic odds ratio (DOR) of 2.37 (95% CI 1.46, 3.82, P=0.0002). Biopsy analysis was successful in 95% of patients within a multisite routine surveillance programme. The probability of dysplasia increased from 11.1% to 17.7% (13.0%, 23.2%) with a positive methylation result consistent with added value in neoplasia detection. To determine added value above colonoscopy alone, a second (reference) colonoscopy was performed in 193 patients without neoplasia. This test also showed an increased number of patients harbouring neoplasia but failed to reach statistical significance (DOR=1.50; 95% CI (0.48, 4.45) P=0.45) The results were also non-significant in the per protocol analysis (DOR=3.93; 95% CI (0.82, 24.75) P=0.09). Patients with persistent abnormal methylation findings at both colonoscopies were at further enhanced risk of neoplasia, 22% of cases (4/18), or 3x that of patients without methylation changes (7/98). ConclusionThis methylation assay was successfully applied within a routine clinical surveillance programme. Blinded analysis confirmed improved rates of neoplasia detection. Although predetermined levels of statistical significance were not reached, the study has also shown that methylation testing can supplement existing clinical and pathological risk stratification, informing patient consent and anticipated dysplasia detection rates. Although not yet recommended for routine uptake, the finding suggest refined methylation assays could be applied for patient benefit.