Optical coherence tomography: a minimally-invasive diagnostic tool for patients with AIN

Abstract

Background: Optical coherence tomography (OCT), a near-infrared imaging technology that produces cross-sectional images of the internal microstructure of biologic tissues, has previously been used to detect squamous cell carcinoma, but not within the anal canal. We hypothesized that OCT could distinguish between patients with anal intraepithelial neoplasia (AIN) and controls. Materials and Methods: We obtained OCT images of the anal canal in 30 patients at high-risk for AIN and 30 controls. High-risk AIN patients underwent high-resolution anoscopy ± biopsy to confirm presence or absence of disease. A blinded analyst measured images for epithelial thickness and brightness. Using criteria from cervical dysplasia literature, 3 blinded and 2 unblinded colorectal surgeons with no prior OCT experience evaluated images for presence of AIN. Sensitivity, specificity, positive predictive value, negative predictive value and kappa statistic, a measure of intra-rater agreement, were calculated. Results: OCT was well-tolerated by all patients. AIN patients had a higher prevalence of immunosuppression (52% vs 13%, p=0.001). Epithelial thickness and brightness were not significantly different between groups. Sensitivity for blinded investigators was 8-31% with 85-78% specificity. Sensitivity for unblinded investigators was 31-77% with 71-75% specificity with a kappa statistic of 0.11. Conclusions: Interpretation of OCT images by naïve observers had better sensitivity, specificity, and negative predictive value when pre-test probability was high. This may improve with observer experience and advancements in technology. Further characterization of the anal canal with newer models for OCT imaging is needed to determine the utility of OCT as a non-invasive clinical adjunct for detection of AIN.