Dynamic optical coherence tomography for cell analysis [Invited]

Abstract

Label-free live optical imaging of dynamic cellular and subcellular features has been made possible in recent years thanks to the advances made in optical imaging techniques, including dynamic optical coherence tomography (D-OCT) methods. These techniques analyze the temporal fluctuations of an optical signal associated with the active movements of intracellular organelles to obtain an ensemble metric recapitulating the motility and metabolic state of cells. They hence enable visualization of cells within compact, static environments and evaluate their physiology. These emerging microscopies show promise, in particular for the three-dimensional evaluation of live tissue samples such as freshly excised biopsies and 3D cell cultures. In this review, we compare the various techniques used for dynamic OCT. We give an overview of the range of applications currently being explored and discuss the future outlook and opportunities for the field.

Fig. 1.

i) schematic showing the usual setups employed in scanning FD OCT vs TD FF-OCT ii) In scanning FD-OCT, one A-scan will be acquired in one shot, carrying the information of the whole depth in z, which will be then extracted in post processing by Fourier-transforming the signal. Subsequently, a B- and C-scans are needed to have the complete view of the three-dimensional sample. In the case of dynamic imaging, two different procedures can be followed: either one can acquire an entire cross-sectional plane first and then follow the evolution of the entire plane over time (BM-scan) or follow the evolution of the intensity over a fixed amount of time for one spot and then shift along the x direction to do the same with all of the other points (MB-scan) iii) For D-FF-OCT the temporal evolution of one transverse plane is recorded and then the focus is shifted in the axial direction to acquire the volume. NPBS: non polarizing beam splitter; Pos.: position; CE: Compensation element, Im. L: imaging lens; FPI: focal plane of interest.

Fig. 2.

Some D-OCT based cell dynamics metrics i) results for motility index (M) and power spectral density (PSD) decay rate (α) of mammary epithelial cells (MEC) spheroids before (blue) and after (red) fixation. As it can be seen, those parameters are significantly different depending on the spheroid condition: the motility is higher for the live sample, the exponential is indeed lower, as expected. Figure taken from Ref. [51] ii) temporal trend of the static power ratio (PR) in gelatin phantoms inseminated with (red curve) and without (green curve) yeast taken from Ref. [23]. The higher activity of the yeast makes the PR decrease faster iii) differences in the time evolution of the logarithmic intensity variance (LIV)(7), and the OCT correlation decay speed (OCDS_e and OCDS_l), respectively, in areas with high and low dynamics. x-axis shows time, y-axis shows the ratio between the area of low/high dynamics and the total area of the sample. Figure taken from Ref. [55]

Fig. 3.

i) H&E-stained histology and scanning dmOCT (RGB method) image of a mouse tongue. The numbers (I-V) indicate the different layer of the tissue, identifiable in both the images: (I) cornified layer, (II) granular and spinous layers, (III) basal layer, (IV) lamina propria and (V) muscle. ( $\ast$ ) indicates cells nuclei. Images taken from Ref. [35] ii) H&E-stained histology and scanning D-µOCT (RGB method) image of a human esophageal biopsy. Cell cytoplasm (green), nuclei (red dots inside) and perinuclear region (blue) can be observed in the µOCT image and they can be found in the histology as well. Frequency bandwidths and corresponding color channels are shown. Images taken from Ref. [11] iii) H&E and dynamic cell imaging (DCI) images of a benign breast lobule (RGB method), double layered structure is identifiable in both the frames. Images taken from Ref. [43]

Fig. 4.

i) Immunohistochemistry (IHC) and D-FF-OCT (HSB method) images of a retinal organoid. Multipotent retinal progenitor cells are tagged with Visual System Homeobox 2 (VSX2) transcription factor, depicted in green. The formation of a rosette is visible in both the images and it is highlighted by the red dotted line. Images are taken from Ref. [12] ii) Cross-sectional OCT images of a MCF-7 spheroid. “Static” image shown in the top left corner and dynamic (HSV method) images shown elsewhere, comparing what can be obtained when the Hue channel is assigned to LIV, OCDS_l and OCDS_e. Images are taken from Ref. [55] iii) Glyph representation of MEC organoids embedded in fibroblasts. The graphical meaning of the three parameters is explained on the bottom. Image taken from Ref. [51] iv) Mosaic of 5 × 5 tiles HSV renders of human fibroblasts, on the right a zoom of the red square area. Actin filaments highlighted with the purple arrow, ring-like subcellular structures pointed at by the orange arrow. Pictures taken from Ref. [61]. Scale bars stand for 40 µm in Fig. 4(iii), 100 µm in Fig. 4(iv) (left panel), 10 µm in Fig. 4(iv) (right panel)