Beyond endoscopic assessment in inflammatory bowel disease: real-time histology of disease activity by non-linear multimodal imaging

Assessing disease activity is a prerequisite for an adequate treatment of inflammatory bowel diseases (IBD) such as Crohn's disease and ulcerative colitis. In addition to endoscopic mucosal healing, histologic remission poses a promising end-point of IBD therapy. However, evaluating histological remission harbors the risk for complications due to the acquisition of biopsies and results in a delay of diagnosis because of tissue processing procedures. In this regard, non-linear multimodal imaging techniques might serve as an unparalleled technique that allows the real-time evaluation of microscopic IBD activity in the endoscopy unit. In this study, tissue sections were investigated using the non-linear multimodal microscopy combination of coherent anti-Stokes Raman scattering (CARS), two-photon excited auto fluorescence (TPEF) and second-harmonic generation (SHG). After the measurement a gold-standard assessment of histological indexes was carried out based on a conventional H&E stain. Subsequently, various geometry and intensity related features were extracted from the multimodal images. An optimized feature set was utilized to predict histological index levels based on a linear classifier. Based on the automated prediction, the diagnosis time interval is decreased. Therefore, non-linear multimodal imaging may provide a real-time diagnosis of IBD activity suited to assist clinical decision making within the endoscopy unit.