Open Access
Add to BibSonomy
Abstract
Scattering-based light sheet microscopy (sLSM) is a microscopy technique that can visualize cellular morphologic details based on the scattering signal. While sLSM was previously shown to image animal tissues ex vivo at a cellular resolution, the wavelength used was chosen based on other in vivo microscopy technologies rather than through a comparison of the sLSM imaging performance between different wavelengths. In this paper, we report the development of a multi-wavelength sLSM setup that facilitates the investigation of different wavelengths for sLSM imaging. Preliminary results of imaging human anal tissues ex vivo showed that the sLSM setup allowed for comparisons of the cellular imaging performance at the same tissue location between different wavelengths. Both the quantitative analysis of the image contrast and the visual assessment by a pathologist showed that the imaging depth increased with wavelength, and the imaging depth increase was most notable around 600 nm. The preliminary results showed that the multi-wavelength sLSM setup could be useful in identifying the optimal wavelength for the specific tissue type.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
In vivo microscopy allows for direct examination of disease-associated cellular morphologic changes from the human tissue without removing the tissue from the patient [1]. Several in vivo microscopy technologies have been translated into clinically-viable medical devices in various clinical fields, including reflectance confocal microscopy in ophthalmology, dermatology, and gastroenterology [2–4], and optical coherence tomography (OCT) in ophthalmology, cardiology, and gastroenterology [5–7]. However, most of the existing in vivo microscopy technologies have challenges in simultaneously achieving a high resolution and a large field of view (FOV). For example, the commercial reflectance confocal microscopes for skin imaging have a high lateral (1.25 µm) and axial resolution (5 µm) [8] thanks to the use of a high-numerical aperture (NA) objective lens. But its FOV is limited to 500–750 µm [9,10], which poses challenges in imaging the entire suspicious tissue region. On the other hand, OCT has a larger FOV, approximately 2–3 mm, but its resolution is around 10–20 µm, which is suitable for imaging architectural features rather than cellular and sub-cellular features [11,12]. High-resolution, large-FOV OCT technologies have been recently developed [13,14]. However, the use of an expensive broadband coherent light source likely increases the device cost.
Light sheet microscopy (LSM) is a microscopy technique that became popular in basic life science research [15,16]. LSM uses separate optical paths for illumination and detection, where the lateral resolution is determined by the detection optics, and the axial resolution by both the illumination and detection optics. Most of the previous LSM work was aimed to achieve a sub-cellular resolution, < 1 µm [16]. If the requirement for the LSM axial resolution could be relaxed to provide an axial resolution used for reflectance confocal microscopy, ∼5 µm, an illumination optics with a low NA (< 0.1) could be used. The low illumination NA in turn could generate a light sheet over a relatively large depth range, hundreds of µm. If the requirement for the lateral resolution of LSM could be relaxed to achieve a lateral resolution used for reflectance confocal microscopy, 1-2 µm, a detection objective lens with a moderate NA, ∼ 0.3 could be used. The moderate-NA objective lens in turn would provide a FOV of several mm, comparable to a typical FOV of OCT.
Taking this approach of achieving a resolution comparable to the resolution of reflectance confocal microscopy while providing a large FOV, we recently demonstrated scattering-based light sheet microscopy (sLSM) of thick, unstained tissues [17]. sLSM detects scattered light signals generated by the refractive index difference between certain sub-cellular/cellular components and their surroundings. In fact, the first LSM setup was based on scattering contrast [18], and several groups successfully demonstrated sLSM imaging of plant roots [19], zebrafish embryos and tumor spheroids [20], and fibroblasts [21]. In our previous experiment, sLSM was shown to visualize cellular details of thick, unstained animal tissues ex vivo with a high resolution (1.8 µm and 6.7 µm for the lateral and axial resolution, respectively) over a large FOV (∼2.5 mm). Another potential advantage of sLSM is that the device cost can be low due to the use of a moderate-NA objective lens, an inexpensive light source such as a LED and superluminescent diode, and a standard CMOS sensor rather than an expensive scientific CMOS sensor.
The previous sLSM setup used a near-infrared spectrum around 834 nm because of its common use in reflectance confocal microscopy and OCT [22,23]. However, there is a need to investigate the optimal wavelength for the specific tissue imaging application. For instance, a shorter wavelength can be used to achieve the same resolution with a lower NA than a longer wavelength. A low illumination NA can increase the light sheet length, and a low detection NA can make the detection objective lens design easier [24]. On the other hand, a longer wavelength is scattered less and therefore can provide a larger imaging depth [25,26]. An sLSM setup that can image the same tissue location with several different wavelengths would facilitate the investigation of the optimal wavelength for the specific tissue type. Such a multi-wavelength sLSM setup needs to be carefully designed to allow for easy changes of the wavelength especially when each lens of the sLSM setup has its own chromatic focal shift.
In this paper, we report the development of a multi-wavelength sLSM setup that facilitates the investigation of different wavelengths for sLSM imaging. We also developed a custom resolution target that enables easy, quantitative measurement of the lateral and axial resolution for each wavelength as a function of imaging depth. Finally, we tested the sLSM setup for imaging human tissues ex vivo with different wavelengths, and the imaging performance was evaluated quantitatively with an edge contrast metric and visually by a pathologist.
2. Method
2.1 Multi-wavelength sLSM setup
The schematic of the multi-wavelength sLSM setup is shown in Fig. 1. In the illumination path, light from a supercontinuum laser (SC-5, YSL Photonics; wavelength range = 470–2400 nm) was coupled into a single-mode fiber (P3-630A-FC-1, Thorlabs; core diameter = 9.0 µm; NA = 0.1 - 0.14). Light from the other end of the fiber was collimated by an achromatic doublet (f = 40 mm). The collimated beam was filtered by a bandpass filter to generate a narrow spectral band (full-width-half maximum, FWHM = 40 nm). Different bandpass filters were used to change the center wavelength from 500 nm to 800 nm with a 100 nm interval. The illumination light was further transmitted through a 3D-printed rectangular aperture and focused by a cylindrical lens (f = 75 mm) and an objective lens (MRH07120, CFI60 Plan Fluor 10x, Nikon; f = 20 mm; NA = 0.3; water immersion) to form a light sheet on the tissue. The width of the aperture along the x-axis was 6.8 mm, which resulted in the light sheet width of 1.68 mm on the tissue. The height of the aperture along the z-axis was adjusted for each wavelength to generate the light sheet thickness of 5 µm (e.g., the aperture height was 1.5 mm for 500 nm, and 1.8 mm for 600 nm). A uniform beam profile was assumed in this theoretical calculation.
The chromatic focal shift of the objective lens was experimentally measured, and the measured chromatic focal shift was 38.4 µm for the spectral range of 500–800 nm. An apochromat objective lens might provide a smaller chromatic focal shift than the current fluorite objective lens. But apochromat, water-dipping objective lenses are not available with the long focal length used in this paper, 20 mm, which is important in achieving a large FOV. The focal length of the objective lens increased as a function of wavelength. If a collimation lens were selected without careful consideration of its own chromatic focal shift, the overall chromatic focal shift could be increased, making it challenging to make the light sheets of different wavelengths overlap with each other. For instance, if an achromatic doublet designed for the visible and near-infrared spectra (49354, Edmund) were used as the collimation lens, it would result in the overall chromatic focal shift of 102 µm. We evaluated several off-the-shelf achromatic doublets and triplets using OpticsStudio (ZEMAX) to find the lens that best compensates for the chromatic focal shift of the objective lens. The chosen achromatic doublet (AC254-040-C, Thorlabs) reduced the overall chromatic focal shift to 24.6 µm.
In the detection path, scattered light from the tissue was collected by a second objective lens (MRH07120, CFI60 Plan Fluor 10x, Nikon) and focused by a camera lens (f = 85 mm; F/2) onto a monochromatic CMOS sensor (acA4024-29um, Basler; 4,024 × 3,036 pixels; pixel size = 1.85 µm). A camera lens rather than a tube lens or achromatic doublet was used because tube lenses typically have a long focal length (180-200 mm), which would generate an image larger than the sensor’s active area, and camera lenses generally have a better correction for field curvature than achromatic doublets with the same focal length. The detection optics had a magnification of 4.25, resulting in the pixel size of 0.44 µm and the FOV of 1.75 mm × 1.32 mm on the tissue xy-plane. The chromatic focal shift of the objective lens, 38.4 µm, was difficult to reduce to a level smaller than the depth of focus (DOF) of the objective lens, 7.4–11.8 µm, using an off-the-shelf camera lens. Therefore, when the wavelength was changed, the focus of the detection objective lens was manually adjusted based on the image sharpness.
2.2 Custom resolution target
The lateral resolution can be measured by imaging standard resolution targets with sharp edges (e.g., USAF resolution target), and the axial resolution can be measured by imaging a mirror. This standard approach, however, requires changes of the specimen between the resolution target and mirror. Microspheres can be used to measure both the lateral and axial resolution, but their distribution is not always uniform, and this approach requires acquisition of volumetric image data with a precise translation of the specimen. When evaluating the multi-wavelength sLSM setup, multiple resolution measurements are required at different depths and for different wavelengths. We developed a custom resolution target (Fig. 2) that allows for direct measurement of the lateral (x-axis) and axial (y-axis) resolution from a single sLSM image.
Fig. 2. Schematic of the custom resolution target and its use for measuring the lateral and axial resolution of sLSM.
The custom resolution target was made of a periodic patten of thin reflective chrome lines (width = 500 nm, period = 30 µm). The target was placed horizontally with the line length parallel to the u-axis. Only a small portion of the target was illuminated by the light sheet, and the illumination width was determined by the light sheet thickness or the axial resolution of sLSM. The portion of the chrome lines within the illuminated region reflected light towards the detection objective lens. The image of the custom resolution target visualized multiple partial lines. The height of each partial line was mainly determined by the light sheet thickness or the axial resolution of sLSM. The width of each partial line image was primarily determined by the line-spread function or lateral resolution of the detection optics. Therefore, both the lateral and axial resolution was measured from a single sLSM image at a given imaging depth. In order to evaluate the resolution at a different depth, the resolution target was translated along the v-axis, which in turn changed the location of the illuminated area on the resolution target. The width of the line pattern, 500 nm, was chosen to keep the measurement error for the lateral resolution small. Using Matlab (Mathworks), we conducted simulation of the measured resolution by convolving the line-spread function of the detection optics with the width of the line pattern and calculating the FWHM of the convolved profile. The simulation results showed that a 10-14% error was expected when measuring the lateral resolution of 0.85 - 1.36 µm, the resolution expected for 0.3 NA and the wavelength of 500–800 nm. The period of the line pattern, 30 µm, was set significantly larger than the line width so that the image of each line does not affect the image of neighboring lines.
The custom resolution target was fabricated by photolithography and etching (Front Range Photomask). The line pattern was made on a glass substrate (thickness = 1 mm) with a chrome coating (optical density = 5). The line pattern width and etching process were iteratively optimized to generate the target line width of 500 nm. The custom resolution target was examined by scanning electron microscopy (Fig. 3(a)) and white light interferometry (Fig. 3(b)). The line width measured by electron microscopy, ∼560 nm, closely matched the design width, 500 nm. There was a broader dark area around the thin line in the electron microscopy image with a width of ∼3 µm. The white light interferometry image (Fig. 3(b)) confirmed that the dark area on the electron microscopy image was not optically reflective, and the strong reflectance signal was confined to the thin line region with a width of ∼560 nm.
Fig. 3. Images of the custom resolution target obtained with scanning electron microscopy (A) and white light interferometry (B).
2.3 Performance test
The axial and lateral resolution was measured using the custom resolution target. Images of the resolution target were acquired, while the resolution target was translated vertically (along the v-axis) with a step size of 10 µm. In each image, the partial lines with a sufficient signal level were analyzed to measure the axial and lateral resolution. When the wavelength was changed, the detection objective focus was manually adjusted to match the detection focal plane to the illumination light sheet.
Tissue imaging performance was evaluated by imaging formalin-fixed human anal tissues. The ex vivo tissue imaging protocol was reviewed and approved by the Stanford Internal Review Board. Index-matching gel (GenTeal Tears Gel, Alcon) was placed on the tissue surface, and a transparent plastic film (material = fluorinated ethylene propylene; refractive index = 1.34; thickness = 250 µm) was placed on the tissue surface. The tissue was then placed under the sLSM setup, and additional index-matching gel was placed between the plastic film and the two objective lenses. The tissue surface was located 80 µm above the illumination focal point of 800 nm along the y-axis so that the illumination DOF for 800 nm, +/- 80 µm, was used to image the superficial region of the tissue. For each wavelength, twenty-five sLSM images were acquired, while the tissue was translated along the u-axis with a step size of 10 µm. The exposure time of the CMOS sensor was set as 1 second. The illumination power on the tissue was 0.19, 0.57, 0.91, and 1.51 µW for 500, 600, 700, and 800 nm, respectively. The gain of the CMOS sensor was adaptively adjusted to provide a similar intensity level for all wavelengths.
2.4 Imaging depth analysis
The imaging depth was evaluated by i) using an edge contrast metric and ii) visual assessment by a pathologist. Since images of the squamous mucosa exhibited multiple layers of squamous cells throughout the epithelium, these images were used for the imaging depth analysis. A flowchart of the quantitative imaging depth analysis is shown in Fig. 4. Each image was convolved with a Gaussian filter with sigma of 5 pixels to reduce the speckle noise while maintaining the image contrast for epithelial cells. The Gaussian-filtered image was convolved with a Laplacian filter with a kernel size of 1(x) × 21(y) pixels. The kernel size of 21 pixels was used because a typical cell-to-cell distance was ∼9 µm or ∼20 pixels. The Laplacian filter was applied along the y-axis to disregard the intensity variation along the x-axis caused by shadowing effects (bright particles on the tissue surface reducing the intensity below). The Laplacian-filtered image was normalized by a smoothed image generated by convolution of the Gaussian-filtered image with an ones vector with a kernel size of 1(x) × 21(y) pixels. The resulting image visualized edges along the vertical direction. The edge-enhanced image was further divided into tall segments with a width of 100 pixels along the x-axis. For each segment, the edge contrast values were averaged along the x-axis at each y coordinate, which generated a curve showing the edge contrast as a function of depth. The contrast curve was fitted with an exponential decay function, e-y/α, with α being the decay factor in the unit of µm. The average and standard deviation of the decay factor were calculated for each wavelength.
Visual assessment of the imaging depth was performed by a pathologist (EY). Each sLSM image was presented to the pathologist without any information about the wavelength used. For each wavelength, five sLSM images obtained from different tissue locations were used for the analysis. The images were presented to the pathologist in a random order for the wavelength and tissue location. Before being presented to the pathologist, the speckle noise of sLSM images was reduced by using a median filter in ImageJ (kernel radius = 2). Ten tick marks equally spaced along the x-axis were imposed on the top of each sLSM image, and the pathologist marked the deepest location where cellular details were observable at each tick marked x location. This resulted in 50 locations examined for each wavelength. The average and standard deviation of the visually-assessed imaging depth were calculated for each wavelength.
3. Results
3.1 Resolution measurement
The measured lateral and axial resolution for different wavelengths is shown in Fig. 5. The measured lateral resolution (Fig. 5(a)) was 1.06 µm, 1.20 µm, 1.37 µm, and 1.58 µm, for 500 nm, 600nm, 700nm, and 800nm, respectively. The lateral resolution was better for shorter wavelengths as expected. The difference between the measured and theoretical lateral resolution was small, 0.18-0.22 µm. The plot for the measured axial resolution (Fig. 5(b)) shows that the foci of the four wavelengths well coincide with each other. When the focus of the 800 nm light sheet was used as the reference, the measured axial resolution for 800 nm was 6.09 µm on average over the theoretical DOF of ± 80 µm. The measured axial resolution was 5.04 µm, 5.55 µm, and 5.90 µm for 500 nm, 600 nm, and 700 nm over their respective theoretical DOF (± 128 µm, ± 107 µm, and ± 91 µm). The measured axial resolution was slightly larger than the target axial resolution, 5 µm, because of the dimensional errors in the 3D-printed apertures and aberrations of the illumination optics.
Fig. 5. Theoretical and measured lateral resolution as a function of wavelength (A) and measured axial resolution for different wavelengths as a function of y coordinate (B).
3.2 Tissue imaging results
Figure 6 shows representative sLSM and histologic images of the anal squamous mucosa. The sLSM images were processed for speckle noise reduction, by applying a gamma value of 0.7 and a median filter with a kernel radius of 2 pixels in ImageJ. In the superficial region of the sLSM images (epithelium or EP in Figs. 6(a)-(d)), layers of stratified squamous epithelial cells are visualized with hypo-reflective (dark) nuclei and hyper-reflective (bright) cytoplasm and cell membranes in a regular honeycomb pattern. The cellular features visualized in the sLSM images were similar to those shown in high-resolution (lateral resolution = 2 µm; axial resolution = 1 µm) OCT images of the cervical epithelium [27], which is known to be histologically similar to the anal epithelium. The cell-to-cell distance in the sLSM images (insets, Figs. 6(a)-(d)) was similar to the distance shown in the histologic image of the same tissue (inset, Fig. 6(e)). Speckle noise is more noticeable in the sLSM images obtained with longer wavelengths, because the same spectral band of 40 nm was used for all the wavelengths, and the spectral averaging of the speckle noise was less effective for longer wavelengths. In deeper regions (lamina propria or LP in Figs. 6(a)-(d)), large and blurry hypo-reflective areas are noticed (arrows in Figs. 6(a)-(d)), which appear to correspond to blood vessels shown in the histologic image (arrow in Fig. 6(e)). The blood vessels were more easily observable in the sLSM images obtained with longer wavelengths.
Fig. 6. sLSM (A-D) and histologic image (E) of the human anal squamous mucosa. EP – epithelium; LP – lamina propria; and arrows - blood vessel.
Figure 7 shows representative sLSM and histologic images of the anal columnar mucosa. The columnar epithelium (EP) was visualized as a hypo-reflective layer in sLSM images (Figs. 7(a)-(d)), which is consistent with the previous reflectance confocal microscopy studies of visualizing goblet cells in the columnar epithelium as hypo-reflective voids [28,29]. Crypts (asterisks) were visualized as dark openings in sLSM images. The junction between the epithelium and lamina propria exhibited a strong signal (arrows in Fig. 7(a)-(d)) in sLSM images, probably due to the light scattering by the basement membrane (arrow in Fig. 7(e)).
Fig. 7. sLSM (A-D) and histologic image (E) of the human anal columnar mucosa. asterisks – crypts; and arrows - junctions between the epithelium and lamina propria.
3.3 Analysis of imaging depth
Representative image contrast curves as a function of imaging depth for different wavelengths are shown in Fig. 8(a). The image contrast decreased as the imaging depth increased. Longer wavelengths showed a slower decrease of the image contrast. The image contrast was higher for longer wavelengths even in superficial imaging depths (left side of Fig. 8(a) curves). Multiply-scattered photons in shorter wavelengths can increase the background signal even at a shallow depth, which is noticeable as a reduced contrast between the hypo- and hyper- reflective regions in Fig. 6(a). The decay factors for different wavelengths are shown in Fig. 8(b). The decay factor increased as a function of wavelength, and the increase was most notable between 500 and 600 nm.
A representative sLSM image manually marked by the pathologist for the deepest imaging depth is shown in Fig. 9(a). The imaging depth and the corresponding tissue surface are marked with red dots. The visually-assessed imaging depth (Fig. 9(b)) showed a similar trend to the quantitatively-calculated decay factor (Fig. 8(b)): the imaging depth increased as a function of wavelength. The visual assessment also showed that an imaging depth of 165–208 µm is useable for examining cellular details of the squamous epithelium.
Fig. 9. Representative sLSM image with red markings for the tissue surface and the deepest tissue depth with visible cellular features (A) and visually-assessed imaging depths for different wavelengths (B).
4. Discussion
In summary, we developed a multi-wavelength sLSM setup to evaluate the sLSM imaging performance for different wavelengths. While we still needed to adjust the detection focus, such adjustment was not needed for the illumination optics thanks to the careful selection of the collimation lens. This facilitated imaging of the same tissue with different wavelengths, and enabled reliable comparisons of different wavelengths. An expensive, supercontinuum laser was used in the current sLSM setup because it delivered different wavelengths through a small-core fiber. However, once the optimal wavelength is determined for the specific tissue imaging application, a clinically-viable sLSM device can be developed using a much cheaper light source such as an LED. A relatively long exposure time, 1 second, was used due to the low illumination power, <1.5 µW. The exposure time can be significantly reduced in future sLSM devices with a single spectral band by optimizing the illumination optics, which can enable real-time imaging, as demonstrated in previous sLSM work [17].
One of the interesting findings was that cellular morphologic details can be visualized from formalin-fixed tissues. This is helpful because the same tissue can be used over time to evaluate different imaging conditions and modalities. Furthermore, there is an abundance of fixed tissue with histopathologic abnormalities that are discarded in the pathology lab after clinical evaluation is complete. This is a rich source of tissue samples that are stable and readily available to the broader scientific community for optical image analysis. However, there is a caveat: fixed tissues might scatter differently than fresh tissues. A future study is needed for analyzing the scattering signal and the visualized cellular details of tissues before and after formalin fixation.
In sLSM images of the columnar mucosa, details of the columnar epithelial cells were not well visualized. This is because the homogenous internal structure of goblet cells did not generate a strong scattering signal. sLSM can be still useful in differentiating between the squamous mucosa and columnar mucosa, and in identifying the transformation zone.
Structures in the lamina propria such as blood vessels and crypts were visualized as hypo-reflective voids. However, the details of these structures were not clear due to the increased light sheet thickness and the resolution degradation at deeper tissue regions. While the capability to discern hypo-reflective structures in the lamina propria even with a degraded resolution could be potentially useful in identifying tumor nodules, its feasibility and utility need to be evaluated in future studies.
Longer wavelengths provided a larger imaging depth based on both the quantitative image contrast evaluation and the visual assessment by a pathologist. This is not surprising because longer wavelengths experience less scattering. However, the amount of increase in the imaging depth was most notable between 500 and 600 nm, and the increase was more moderate after 600 nm (Figs. 8 and 9). The visual assessment results show that 600 nm allows for the visualization of cellular details up to a depth of ∼200 µm, which can be useful in evaluating epithelial cellular morphologic changes in malignant diseases. Therefore, a wavelength around 600 nm can be used for further development of a clinically-viable sLSM device. The use of 600 nm, rather than a near-infrared wavelength around 800 nm, could either increase the light sheet length for the same axial resolution or improve the axial resolution for the same light sheet length. However, further studies will be needed for imaging fresh human tissues with different wavelengths to determine the optimal wavelength for the specific application.
Funding
National Institute of Biomedical Imaging and Bioengineering (R21EB030079).
Disclosures
The University of Arizona has a technology-licensing agreement with ArgosMD on the portable confocal microscopy technology. DK has the rights to receive royalties as a result of this licensing agreement. DK serves as a scientific advisor to ArgosMD.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
1. W. A. Wells, M. Thrall, A. Sorokina, J. Fine, S. Krishnamurthy, A. Haroon, B. Rao, M. M. Shevchuk, H. C. Wolfsen, G. J. Tearney, and L. P. Hariri, “In vivo and ex vivo microscopy: moving toward the integration of optical imaging technologies into pathology practice,” Arch. Pathol. Lab. Med. 143(3), 288–298 (2019). [CrossRef]
2. M. Tavakoli, P. Hossain, and R. A. Malik, “Clinical applications of corneal confocal microscopy,” Clin. Ophthalmol. 2(2), 435–445 (2008).
3. S. González, K. Swindells, M. Rajadhyaksha, and A. Torres, “Changing paradigms in dermatology: confocal microscopy in clinical and surgical dermatology,” Clin. Dermatol. 21(5), 359–369 (2003). [CrossRef]
4. R. Kiesslich, M. Goetz, M. Vieth, P. R. Galle, and M. F. Neurath, “Confocal laser endomicroscopy,” Gastrointest. Endosc. Clin. N. Am. 15(4), 715–731 (2005). [CrossRef]
5. W. Drexler and J. G. Fujimoto, “State-of-the-art retinal optical coherence tomography,” Prog. Retin. Eye Res. 27(1), 45–88 (2008). [CrossRef]
6. T.-H. Tsai, C. L. Leggett, A. J. Trindade, A. Sethi, A.-F. Swager, V. Joshi, J. J. Bergman, H. Mashimo, N. S. Nishioka, and E. Namati, “Optical coherence tomography in gastroenterology: a review and future outlook,” J. Biomed. Opt. 22(12), 1–17 (2017). [CrossRef]
7. T. Yonetsu, B. E. Bouma, K. Kato, J. G. Fujimoto, and I.-K. Jang, “Optical coherence tomography,” Circ. J. advpub 2(1-2), 9–25 (2013). [CrossRef]
8. E. Cinotti, L. Gergelé, J. L. Perrot, A. Dominé, B. Labeille, P. Borelli, and F. Cambazard, “Quantification of capillary blood cell flow using reflectance confocal microscopy,” Ski. Res. Technol. 20(3), 373–378 (2014). [CrossRef]
9. L. E. Meyer, N. Otberg, W. Sterry, and J. Lademann, “In vivo confocal scanning laser microscopy: comparison of the reflectance and fluorescence mode by imaging human skin,” J. Biomed. Opt. 11(4), 044012 (2006). [CrossRef]
10. G. Peterson, D. K. Zanoni, M. Ardigo, J. C. Migliacci, S. G. Patel, and M. Rajadhyaksha, “Feasibility of a video-mosaicking approach to extend the field-of-view for reflectance confocal microscopy in the oral cavity in vivo,” Lasers Surg. Med. 51(5), 439–451 (2019). [CrossRef]
11. J. Welzel, “Optical coherence tomography in dermatology: a review,” Ski. Res. Technol. 7(1), 1–9 (2001). [CrossRef]
12. R. A. Leitgeb, “En face optical coherence tomography: a technology review [Invited],” Biomed. Opt. Express 10(5), 2177–2201 (2019). [CrossRef]
13. L. Liu, J. A. Gardecki, S. K. Nadkarni, J. D. Toussaint, Y. Yagi, B. E. Bouma, and G. J. Tearney, “Imaging the subcellular structure of human coronary atherosclerosis using micro–optical coherence tomography,” Nat. Med. 17(8), 1010–1014 (2011). [CrossRef]
14. A. Dubois, O. Levecq, H. Azimani, D. Siret, A. Barut, M. Suppa, V. del Marmol, J. Malvehy, E. Cinotti, P. Rubegni, and J.-L. Perrot, “Line-field confocal optical coherence tomography for high-resolution noninvasive imaging of skin tumors,” J. Biomed. Opt. 23(10), 1–9 (2018). [CrossRef]
15. R. M. Power and J. Huisken, “A guide to light-sheet fluorescence microscopy for multiscale imaging,” Nat. Methods 14(4), 360–373 (2017). [CrossRef]
16. J. M. Girkin and M. T. Carvalho, “The light-sheet microscopy revolution,” J. Opt. 20(5), 053002 (2018). [CrossRef]
17. C. D. Nguyen, P. K. O’Neal, N. Kulkarni, E. Yang, and D. Kang, “Scattering-based light-sheet microscopy for rapid cellular imaging of fresh tissue,” Lasers Surg. Med. 53(6), 872–879 (2021). [CrossRef]
18. H. Siedentopf and R. Zsigmondy, “Uber Sichtbarmachung und Größenbestimmung ultramikoskopischer Teilchen, mit besonderer Anwendung auf Goldrubingläser,” Ann. Phys. 315(1), 1–39 (1902). [CrossRef]
19. Z. Yang, H. Downie, E. Rozbicki, L. X. Dupuy, and M. P. MacDonald, “Light Sheet Tomography (LST) for in situ imaging of plant roots,” Opt. Express 21(14), 16239–16247 (2013). [CrossRef]
20. D. Di Battista, D. Merino, G. Zacharakis, P. Loza-Alvarez, and O. E. Olarte, “Enhanced light sheet elastic scattering microscopy by using a supercontinuum laser,” Methods Protoc. 2(3), 57 (2019). [CrossRef]
21. M. Lin, X. Qiao, Q. Liu, C. Shao, and X. Su, “Light-sheet-based 2D light scattering cytometry for label-free characterization of senescent cells,” Biomed. Opt. Express 7(12), 5170–5181 (2016). [CrossRef]
22. K. S. Nehal, D. Gareau, and M. Rajadhyaksha, “Skin imaging with reflectance confocal microscopy,” in Semin. Cutaneous Med. Surg. (WB Saunders, 2008), pp. 37–43.
23. Y. Wang, J. S. Nelson, Z. Chen, B. J. Reiser, R. S. Chuck, and R. S. Windeler, “Optimal wavelength for ultrahigh-resolution optical coherence tomography,” Opt. Express 11(12), 1411–1417 (2003). [CrossRef]
24. Y. Zhang and H. Gross, “Systematic design of microscope objectives. Part I: System review and analysis,” Adv. Opt. Technol. 8(5), 313–347 (2019). [CrossRef]
25. P. Taroni, A. Pifferi, A. Torricelli, D. Comelli, and R. Cubeddu, “In vivo absorption and scattering spectroscopy of biological tissues,” Photochem. Photobiol. Sci. 2(2), 124–129 (2003). [CrossRef]
26. S. L. Jacques, “Optical properties of biological tissues: a review,” Phys. Med. Biol. 58(11), R37–R61 (2013). [CrossRef]
27. H. M. Leung, M. L. Wang, H. Osman, E. Abouei, C. MacAulay, M. Follen, J. A. Gardecki, and G. J. Tearney, “Imaging intracellular motion with dynamic micro-optical coherence tomography,” Biomed. Opt. Express 11(5), 2768–2778 (2020). [CrossRef]
28. D. Kang, M. J. Suter, C. Boudoux, H. Yoo, P. S. Yachimski, W. P. Puricelli, N. S. Nishioka, M. Mino-Kenudson, G. Y. Lauwers, B. E. Bouma, and G. J. Tearney, “Comprehensive imaging of gastroesophageal biopsy samples by spectrally encoded confocal microscopy,” Gastrointest. Endosc. 71(1), 35–43 (2010). [CrossRef]
29. D. Kang, S. C. Schlachter, R. W. Carruth, M. Kim, T. Wu, N. Tabatabaei, A. R. Soomro, C. N. Grant, M. Rosenberg, N. S. Nishioka, and G. J. Tearney, “Large-area spectrally encoded confocal endomicroscopy of the human esophagus in vivo,” Lasers Surg. Med. 49(3), 233–239 (2017). [CrossRef]
