In and , OCT and OCE images of skin from the tip of the middle finger of a 29-year-old male are presented. For both images, the acquisition time was 1.4 s and the number of pixels in the z
-direction is 700. The number of pixels in the x
-direction in is 7,000; determined by the number of A-scans recorded per B-scan, and 175 in ; determined by the number of excitation cycles. Dense sampling in the x-direction was chosen to minimize phase errors due to lateral scanning. To reduce noise, a median filter (7 × 3 kernel) was applied. For the OCE data, it was applied following thresholding and prior to extraction of the strain rate magnitude. The unequal kernel size is to account for different pixel numbers in the x
directions. In , the glass imaging plate, stratum corneum and living epidermis are labeled. The stratum corneum is the lower scattering, more superficial layer of the epidermis and the living epidermis is the higher scattering, deeper region consisting of papillary ridges. Strong contrast is observed between the stratum corneum and the living epidermis. The stratum corneum has a thickness, in optical pathlength, of approximately 300 μm, consistent with previous OCT measurements [32
]. A sweat gland is also present, indicated by higher signal intensity than the surrounding stratum corneum. The OCE image recorded from the same region is presented in . In , the OCE signal (color map) is overlaid on the OCT signal (grayscale). Due to the OCE signal density, only the top 40% of the OCE signal is presented, so that the underlying OCT signal is visible. Contrast between the stratum corneum and the living epidermis is visible in all images. The strain rate magnitude on average is higher in the living epidermis than in the stratum corneum, suggesting it has a more elastic response, as previously reported [12
]. Although there are subtle differences in the two images, in general, the strain rate magnitude tracks proportionally the OCT image signal intensity. We speculate that this is because higher optical scattering areas are more dense [32
] and, therefore, stiffer. It should be noted that higher optical scattering may also be caused by differences in refractive index.
A significant improvement in OCE image resolution is visible in in comparison to in vivo
OCE images previously presented [12
]. This is largely due to the faster acquisition speed obtained by using the SD-OCT system, which results in significantly reduced motion artifacts.
3D-OCT and OCE images of skin from the middle finger of the same subject are presented in
. OCE B-scans at each y
-position were generated using the technique described above. The total acquisition time for a 3D-OCE data set was 5 min for images with dimensions (xyz
) of 2 mm × 1 mm × 1 mm. To reduce motion artifacts, the finger was strapped in position prior to each measurement. To keep the acquisition time as short as possible, only 50 excitation cycles per B-scan were used; reducing the B-scan acquisition time to 0.4 s. This was achieved by reducing the number of A-scans per B-scan from 7,000 to 2,000; resulting in a lateral resolution of 40 μm in the x
-direction. The reduction of spatial resolution is visible in . In the z
-direction, there are 1,024 pixels and in the y
-direction, 100 pixels (1-mm range). The raw OCE and OCT data sets were cropped and transformed within an image processing package [34
] and then normalized so that the intensity range was 0.0-1.0 [35
]. The data sets were then imported into a volume exploration and presentation tool [36
], and reconstructed into 3D data sets at full resolution. The corresponding visualizations were produced from view-aligned, slice-based rendering. Final pixel contributions were defined by applying a two-dimensional transfer function that weighted the opacity and color of each voxel based on the intensity and gradient value in the volumetric data sets.
Fig. 4 3D visualization of in vivo skin from the middle finger of a male subject. (a) OCT, (b) OCE, and (c) overlay, from first perspective view; (d) OCT, (e) OCE, and (f) overlay, from second perspective view; (g) OCT, (h) OCE and (i) overlay, from en face (more ...)
In and , two perspective views of the OCT data are presented, with one corner cut away to a depth of 300 μm, revealing internal structure. In and , the OCE signal is displayed and, in and , the OCE data is overlaid on the OCT data, as in . In these images, the highest OCE signal is visible in the living epidermis and the sweat glands, consistent with the result presented in . In -
views from the surface of the tissue are presented. In the OCT image, presented in , sweat glands are visible as regions of high signal intensity. Large variations in the strain rate magnitude are visible in in the regions corresponding to sweat glands in . Additional regions of high OCE signal in may correspond to sweat glands not visible in the OCT data. The OCE signal overlaid on the OCT signal is presented in . A second en face
view is presented in - at a depth of 300 μm. The regions of high OCT signal at this depth correspond to the living epidermis. Shadows of the sweat glands are visible in . Several of these shadow artifacts are indicated in the figure (white arrows). In the OCE images presented in , the living epidermis is represented by regions of high OCE signal. This is consistent with the result presented in . The OCE signal overlaid on the OCT signal is presented in . The full 3D data sets are also available,
The skin on the middle finger of the same subject was hydrated in warm water for 30 min and then imaged with the same acquisition settings, preload, excitation amplitude and frequency, and median filtering as for the images presented in . An OCT image of the hydrated skin is presented in
. The imaging plate, stratum corneum and living epidermis are readily distinguished. The OCE image is presented in . The contrast between stratum corneum and living epidermis is reduced in comparison to the results presented in . This is attributed to a more elastic response of the stratum corneum in the hydrated case [37
]. An artifact is visible in – the surface of the imaging plate appears not to be perfectly flat. We believe that this is caused by a variation in the thickness of the index-matching glycerol used, which results in a slight variation in the optical path length. As in , in the top 40% of the OCE signal (color map) is overlaid on the OCT signal (grayscale).
(a) OCT; (b); OCE and; (c) overlaid images of the in vivo hydrated skin on the middle finger. Image dimensions are 1.4 mm × 1.4 mm.
3D-OCT and OCE imaging were also performed on the hydrated skin and are presented in
. In the hydrated 3D-OCE case, 250 excitation cycles were introduced across a lateral range of 2 mm, compared with 50 excitation cycles across the same distance in . This resulted in the lateral resolution in the x-direction being determined by the lateral resolution of the OCT system. The trade-off is that the acquisition time increased by a factor of five, resulting in an acquisition time of 25 min, and motion artifacts became more prominent and were manifested by geometrical distortion of skin features.
Fig. 6 In vivo 3D visualizations of hydrated skin from the middle finger of a male subject. (a) OCT, (b) OCE, and (c) overlay, from first perspective view; (d) OCT, (e) OCE, and (f) overlay, from second perspective view; (g) OCT, (h) OCE and (i) overlay, from (more ...)
In a similar manner to the results presented in and , in and perspective views of the OCT data from the hydrated skin are presented, with one corner cut away to a depth of 300 μm, revealing internal structure. In and , the OCE signal is displayed and, in and , the OCE data is overlaid on the OCT data. On average the OCE signal is higher in the hydrated stratum corneum when compared with the unhydrated stratum corneum shown in and , suggesting a more elastic response. In -, en face
views from the surface of the tissue are presented. In comparison to , the hydrated en face
OCT image from the skin surface () appears relatively uniform. The sweat glands are not visible in the hydrated case. However, large variations in the OCE image are visible in . We speculate that this is due to variations in the hydration state of the skin. A second en face
view is presented in - at a depth of 300 μm. Large variations in the OCE signal, possibly due to changes in hydration level, are also visible at this depth. The full 3D data sets are also available,