Fig 1.
Schematic describing the experimental setup for in-vivo multiphoton autofluorescence microscopy (MPAM) and second harmonic generation microscopy (SHGM).
HWP: half wave plate; GT: Glans Thompson prism; BE: beam expander; DF: dichroic filter. The hamster is placed on the stage of the inverted microscope with the buccal pouch exposed and facing the objective.
Fig 2.
Early events during neoplastic transformation.
Schematic representation of the current understanding of cellular and layer-based changes with neoplasia (a-c), MPAM-SHGM micrographs (d-f) and histological sections (g-i) of hamster oral mucosa for normal (a, d, g), mild dysplasia (b, e, h) and severe dysplasia (c, f, i) are shown. In-vivo autofluorescence from MPAM (Red) co-registered with SHG (Green) are overlaid to produce cross-sectional MPAM-SHGM micrographs (d-f) comparable to histology (g-i). K: Keratinizing layer; E: Epithelial layer, M: Lamina propria or ECM. White and black arrows point towards ECTI and “*” shows enlarged nuclei in dysplastic epithelium. Scale bar: 50μm.
Fig 3.
Cytological features visible in in-vivo MPAM are shown from normal (a), and dysplastic (b,c) epithelium of hamster oral mucosa.
White boxes in “a” and “b” shows comparison of nuclear density between a normal (0.25 nuclei/100 μm2) and dysplastic epithelium (0.12 nuclei/100 μm2). “*” shows presence of keratin pearls (b) and arrows point towards enlarged nuclei (c) in dysplastic epithelium.
Fig 4.
Epithelial Thickness measurements from MPAM-SHGM (a) and histology (b) show statistically significant increase in thickness of mild, moderate and severe dysplasia from normal mucosa.
Box plot analyses show the median, 25th and 75th percentiles, and outliers for both MPAM-SHGM and histology. “*” and “**” represent statistical difference between groups at p<0.05 and 0.01 significance respectively.
Fig 5.
Cross-sectional x-z SHGM micrograph is shown in (a). Figure (b) shows an ECTI (solid white line) extracted from (a) after the FeatureJ: Edges plugin in ImageJ was applied.
The dashed line represents the reference linear distance between the two ends of the image. The equations on the right were used to calculate ΔLinearity (ΔLnorm). Scale bar: 50 μm.
Fig 6.
ΔLnorm measurements from MPAM-SHGM (a) and histology (b) show statistically significant increase in ΔLinearity of mild, moderate and severe dysplasia from normal mucosa.
Box plot analyses show the median, 25th and 75th percentiles, and outliers for both MPAM-SHGM and histology. “*” and “**” represent statistical difference between groups at p< 0.05 and 0.01 significance respectively.
Table 1.
Summary of the statistical analyses performed from ΔLinearity measurements carried out on cross-section images from MPAM-SHGM or histopathology.
Fig 7.
ROC curve at 95% confidence for ΔLinearity calculated from MPAM-SHGM (Blue) and histological (Green) images.