Objectives/Goals: Colorectal cancer (CRC) disproportionately affects the U.S. Deep South. Socioeconomics, genetics, and healthcare access contribute to increased incidence and mortality. A next-generation endoscope with improved contrast and automatic detection may boost early detection and narrow expertise gaps between top centers and rural clinics. Methods/Study Population: We implemented excitation-scanning hyperspectral imaging (Ex-HSI) to discriminate normal and cancerous mouse colorectal tissues by detecting changes in autofluorescence. Ex-HSI permits detection of all emitted light above a cutoff wavelength. Ex-HSI reduces acquisition time and improves signal-to-noise ratio compared to traditional emission scanning. We used azoxymethane/dextran sodium sulfate (AOM/DSS) to induce colitis and nodule formation in the mouse CRC model. Ex-HSI data was complemented with transmitted light and confocal images. Histology sectioning with H&E staining was obtained for “gold standard” validation. Results/Anticipated Results: Inflammation and bleeding were observed, consistent with AOM/DSS treatment. Nodules were visible using 4x and 20x objectives. Background-subtracted Ex-HSI data was corrected to a flat spectral response. Excitation spectra were extracted from select regions within each field of view. Excitation spectra displayed multiple peak excitation wavelengths, suggesting multiple autofluorescent molecules present. Further investigation will utilize spectral unmixing, principal component analysis, and neural networks to interpret image data. Discussion/Significance of Impact: A hyperspectral endoscope that detects early-stage cancers and precancers would enhance CRC screening in the Deep South. This research is supported by NIH P01HL066299, NIH UL1TR00141, S10RR027535, S10OD02860, NSF MRI1725937, CCTS T32TR004767, and the Center for Lung Biology, University of South Alabama.
Tang-Holmes et al. (Wed,) studied this question.