Volumetric Multiphoton Autofluorescence Microscopy (MPAM) and Second Harmonic Generation Microscopy (SHGM) show promise for revealing indicators of neoplasia representing the complex microstructural organization of mucosa, potentially providing high specificity for detection of neoplasia, but is limited by small imaging area. Large area fluorescence methods on the other hand show high sensitivity appropriate for screening but are hampered by low specificity. In this study, we apply MPAM-SHGM following guidance from large area fluorescence, by either autofluorescence or a targeted metabolic fluorophore, as a potentially clinically viable approach for detection of oral neoplasia.
Sites of high neoplastic potentially were identified by large area red/green autofluorescence or by a fluorescently labelled deoxy-glucose analog, 2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-D-glucose (2-NBDG) to highlight areas of high glucose uptake across the buccal pouch of a hamster model for OSCC. Follow-up MPAM-SHGM was conducted on regions of interests (ROIs) to assess whether microscopy would reveal microscopic features associated with neoplasia to confirm or exclude large area fluorescence findings. Parameters for analysis included cytologic metrics, 3D epithelial connective tissue interface metrics (MPAM-SHGM) and intensity of fluorescence (widefield). Imaged sites were biopsied and processed for histology and graded by a pathologist. A small sample of human ex vivo tissues were also imaged.
A generalized linear model combining image metrics from large area fluorescence and volumetric MPAM-SHGM indicated the ability to delineate normal and inflammation from neoplasia.