Parametric imaging of attenuation by optical coherence tomography: review of models, methods, and clinical translation

Abstract

Significance: Optical coherence tomography (OCT) provides cross-sectional and volumetric images of backscattering from biological tissue that reveal the tissue morphology. The strength of the scattering, characterized by an attenuation coefficient, represents an alternative and complementary tissue optical property, which can be characterized by parametric imaging of the OCT attenuation coefficient. Over the last 15 years, a multitude of studies have been reported seeking to advance methods to determine the OCT attenuation coefficient and developing them toward clinical applications.

Aim: Our review provides an overview of the main models and methods, their assumptions and applicability, together with a survey of preclinical and clinical demonstrations and their translation potential.

Results: The use of the attenuation coefficient, particularly when presented in the form of parametric en face images, is shown to be applicable in various medical fields. Most studies show the promise of the OCT attenuation coefficient in differentiating between tissues of clinical interest but vary widely in approach.

Conclusions: As a future step, a consensus on the model and method used for the determination of the attenuation coefficient is an important precursor to large-scale studies. With our review, we hope to provide a basis for discussion toward establishing this consensus.

Keywords: attenuation coefficient; cardiology; dermatology; multiple scattering; oncology; optical coherence tomography; single scattering.

Fig. 1

Propagation of a focused OCT beam and ray representation of scattering in the sample.

Fig. 2

Simulated average A-scan $⟨A(z)⟩$ according to Eq. (2) (black curve) and the contribution of the individual terms, including single exponential decay (red dash), CPSF $t(z)$ (blue dash), and sensitivity roll-off $h(z)$ (green dash). Parameters for the simulation are: $z_0=0.25\mathrm{mm}$; ${\mu }_s=5{\mathrm{mm}}^{-1}$; refractive index $n=1.4$; noise floor at $-80\mathrm{dB}$; Rayleigh length $z_R=100\mu \mathrm{m}$; focus location $z_f=1\mathrm{mm}$; center wavelength and spectral sampling increment are $\lambda =1300\mathrm{nm}$ and $\mathrm{\Delta }\lambda =0.1\mathrm{nm}$, respectively, giving a maximum imaging depth of $z_D=3\mathrm{mm}$; $s=2$.

Fig. 3

Procedure of fitting to retrieve the attenuation coefficient. One A-scan in (b) is extracted from an intravascular OCT image in (a) along the red line, speckle smoothed and fitted in subsequent windows. The retrieved ${\mu }_{\mathrm{OCT}}$ is plotted as an overlay on the grayscale image in (c). Colormap: 0 to $12{\mathrm{mm}}^{-1}$.

Fig. 4

Average OCT amplitude for samples of pure silica beads in water. The $0.5\text{-}\mu \mathrm{m}$ beads (black) with scattering coefficient of $5{\mathrm{mm}}^{-1}$ and anisotropy factor of 0.1 and $1.5\text{-}\mu \mathrm{m}$ beads (red) with scattering coefficient of $5{\mathrm{mm}}^{-1}$ and anisotropy factor of 0.9. OCT data were collected using a swept-source 1300-nm system with a 150-mm-focal length detection lens.

Fig. 5

Schematic of OCT parametric attenuation coefficient imaging. The A-scan at each lateral location is averaged in a lateral ($x\text{–}y$) window outlined by the cuboid and used to calculate the attenuation coefficient in a depth ($z$) window, leading to a 2-D map in the en face ($x\text{–}y$) plane. Courtesy of Blake R. Klyen (unpublished).

Fig. 6

OCT attenuation imaging of human skin in vivo. (a) OCT vasculature image of normal skin. (b) Parametric attenuation coefficient imaging of the tissue region in the blue square in (a). Dashed circles outline the regions with incorrect attenuation coefficients due to the blood vessels. AFR is $200\mu \mathrm{m}$ with lateral averaging of $40\times 40\mu \mathrm{m}$. (c) An example showing incorrect fitting caused by vessels, from the zone marked by the lower of the purple squares in (b). (d) and (e) Longitudinal parametric attenuation coefficient imaging of a human burn scar before and after laser treatment with vascular masks shown in black. AFR is $250\mu \mathrm{m}$ with lateral averaging of $20\times 20\mu \mathrm{m}$. Adapted from Refs.  and .

Fig. 7

Schematic of the vessel wall with an advanced atherosclerotic plaque. The healthy wall consists of three layers (top): the intima, lying directly beneath the endothelium, the media, which consists of SMCs, and the adventitia, which is made up of connective tissues. These layers are separated by elastic membranes (not shown). A large thin-cap fibroatheroma (bottom) is a heterogeneous structure that exhibits a number of characteristics that may be recognized in OCT images and affect the attenuation coefficient. Adapted from Ref. .

Fig. 8

OCT attenuation imaging of a coronary artery with an atherosclerotic lesion in vitro. (a) and (b) OCT image and corresponding histology of the artery. (c) Cartoon overlaid on the histology to indicate an advanced necrotic core (red) behind a calcification (gray), and a slight fibrotic (green) circumferential intimal thickening. (d) OCT attenuation coefficient image ranging from 0 (blue) to $15{\mathrm{mm}}^{-1}$ (red). Scale bar: 1 mm. Adapted from Ref. .

Fig. 9

Combined imaging of OCT attenuation and backscattering coefficient of a fibrocalcific plaque. (a) and (b) Images of histology and OCT attenuation coefficient. (c) Image of the combined attenuation and backscattering coefficient using the colormap in (d). C, calcific tissue; F, fibrous tissue; L, lipid tissue. The three tissue types led, respectively, to attenuation coefficients of $5.7\pm 1.4{\mathrm{mm}}^{-1}$, $6.4\pm 1.2{\mathrm{mm}}^{-1}$, and $13.7\pm 4.5{\mathrm{mm}}^{-1}$; and backscattering coefficients of $4.9\pm 1.5{\mathrm{mm}}^{-1}$, $18.6\pm 6.4{\mathrm{mm}}^{-1}$, and $28.2\pm 8.9{\mathrm{mm}}^{-1}$, respectively. Adapted from Ref. .

Fig. 10

OCT attenuation imaging of a malignant human axillary lymph node. (a)–(c) Histology, OCT attenuation, and structural OCT image of the lymph node. The circles highlight the residual, noncancerous cortical tissue. Scale bars: 1 mm. Adapted from Ref. .