Detection and characterisation of dysplasia within the colon

Introduction
Patients with Inflammatory bowel disease (IBD)-colitis have an increased risk of developing colorectal cancer (CRC). CRC associated with IBD-colitis is understood to evolve along the inflammation-dysplasia-cancer pathway. Optimising the detection and characterisation of dysplasia is paramount for delivering effective surveillance, in order to prevent progression of dysplasia to cancer.

Aims
• Are random biopsies during surveillance colonoscopy in patients with IBD colitis still required
• To determine the optimal endoscopic method for the detection of dysplasia in patients with IBD colitis during surveillance and is there a subpopulation in which invisible dysplasia exists
• To determine the accuracy of in-vivo lesion characterisation in real-life practice

Methods
Two systematic reviews/meta-analyses were conducted 1) To determine what proportion of patients with dysplasia during surveillance colonoscopy are identified by random biopsy alone and if the proportion detected is influenced by the cohort’s perceived risk for dysplasia 2) The diagnostic accuracy of optical imaging techniques for in-vivo lesion characterization in colonic IBD. A randomised controlled trial was then performed to determine whether 1) The concentration of dye (0.03% versus 0.2% indigo carmine) impacts on the detection of dysplastic lesions in patients with colitis 2) Understand the optical diagnostic accuracy of in-vivo lesion characterisation for colonic lesions detected 3) Predictors for invisible dysplasia.

Results
On pooling proportions, 13.05% (95% CI 7.28 – 19.87%) of patients with dysplasia had this identified by random biopsies alone. Pooled proportion of patients with dysplasia identified by random biopsy alone within the high-risk group was more than double, 14.19% (95% CI 7.43 – 22.29%), compared to the low-risk group, 6.42 (95% CI 0.04 – 18.45%). The lesion characterisation meta-analysis showed Confocal Laser Endomicroscopy (CLE) as the most accurate technology, with a sensitivity of 87% (95% CI 71%-95%), specificity of 94% (95% CI 87%-97%), area under the SROC curve of 0.96 (95% CI 0.94-0.97).
For the RCT, 300 procedures were randomised into 0.2% (n=150) and 0.03% (n=150) indigo carmine concentrations. Targeted neoplasia was detected in 32 (21.3%) procedures in the 0.2% arm and 26 (17.3%) procedures in the 0.03% arm; p=0.465. Nine procedures (3%) had random biopsy only dysplasia, with 88.9% in the BSG high risk group. Overall sensitivity, specificity, positive and negative predictive values for dysplasia optical diagnosis were 84.7%, 82.3%, 64.3% and 93.5%. Accuracy further improved using the 0.2% dye concentration.

Conclusion
Dye-based Chromoendoscopy has a high yield for detecting targeted dysplasia regardless of the concentration used, although numerically favours the 0.2% indigo carmine using the spray catheter. Accuracy of in-vivo lesion characterisation also favours the more concentrated dye solution. When using high definition chromoendoscopy, random biopsies for invisible dysplasia are not required when patients are risk stratified within the BSG low and intermediate risk groups.