Identifying Colonic Dysplasia and Cancer From Rectal Nuclear Nano-morphology Markers in Ulcerative Colitis Patients

Abstract

Introduction: Current guidelines recommend that ulcerative colitis (UC) pts undergo annual surveillance colonoscopy with random colonic biopsies for detection of colorectal dysplasia and cancer. This is a suboptimal invasive technique for detecting all colorectal dysplasia. Less invasive surveillance measures that would stratify pts based on risk of developing dysplasia/cancer are needed. We hypothesize, based on the premise of cancer field effect, that markers from rectal biopsies can predict the presence of dysplastic/neoplastic lesions in the colon. This pilot study aims to evaluate whether Spatial-domain Low-coherence Quantitative Phase Microscopy (SL-QPM) derived nuclear nano-morphology markers from flat normal appearing rectal mucosa can detect remote colonic dysplasia/neoplasia in UC pts. Methods: We prospectively recruited UC pts undergoing surveillance colonoscopy with random colon biopsies as per the American Gastroenterological Association surveillance guidelines. Two extra biopsies from flat rectal mucosa were obtained, fixed in formalin and processed per standard histology protocol. According to the final histologic diagnoses from all random biopsies, pts were categorized into a lowrisk group, if pathology showed no dysplasia/cancer, or a high-risk group, if there was any dysplasia, including adenomatous tissue or colon adenocarcinoma. For each pt, SL-QPM analysis was performed on nuclei from 40 to 60 rectal columnar epithelial cells.Figure 1Results: A total of 148 pts were included; 127 low risk and 21 high risk. The high risk group included pts with colorectal cancer (n=1), low-grade dysplasia (n=5), high-grade dysplasia (n=1), indefinite dysplasia (n=2) or sporadic tubular adenoma (n=12). A set of nuclear nano-morphology markers distinguished low risk from high risk UC pts with high statistical significance. Results were not confounded by presence of active inflammation anywhere in the colon (p-value>0.5). The performance characterized by area under receiver operating characteristics curve is 0.82 in distinguishing low from high risk UC pts (86% sensitivity at 61% specificity). Conclusion: Nuclear nano-morphology markers from non-dysplastic rectal tissue identify remote colonic dysplasia in UC pts. This technology is unaffected by the presence of inflammation and stratifies patients into low versus high risk for harboring dysplasia/cancer. This approach could enhance the efficiency of dysplasia surveillance in UC patients and an ongoing study is underway to validate these findings.Figure 2Figure 3