Imaging a thin section within
thick living tissue
One of the primary differences between
imaging living tissue and carefully
prepared biopsy slides is dimensionality.
In traditional biopsy, extracted tissue is
embedded in paraffin, sliced into very
thin sections (about a cell-layer thick),
and mounted on a microscope slide
for inspection. This results in a two-dimensional sample that resides within
the depth of field of the microscope on
which it is viewed.
On the other hand, in vivo tissue is
inherently three-dimensional. Imaging
thick tissue samples is a challenge that is
addressed in different ways by confocal
microscopy, nonlinear microscopy and
OCT. The confocal microendoscope
employs the confocal principle, whereby
the co-registration of an illumination
focal point and the detection sensitivity
distribution of a confocal aperture (
pinhole) produces a net sensitivity function
that is highly localized in space.
With two-dimensional scanning of the
confocal point across the sample, a high-resolution image is obtained of a thin
two-dimensional “optical section” within
three-dimensional tissue. Typically, scanning is in the x-y plane perpendicular to
the optical axis at a fixed depth relative
to the tissue surface. The thickness of the
optical section is usually on the order of
a single cell layer or less ( 1-20 m) and
depends on the numerical aperture of the
illumination and collection beams.
Nonlinear microscopy is similar
except that it uses a high-power short-pulse laser source whose instantaneous
irradiance is sufficient to excite a
nonlinear response (e.g., two-photon
fluorescence) within a very small focal
volume. Again, by scanning the illumination point in the x-y plane across
the sample and detecting the emitted
nonlinear signal, one can build up a
high-resolution two-dimensional image
of a thin section of tissue. The section
thickness is typically a few microns, and
the in-plane resolution is high because
a high-numerical-aperture beam is used
to achieve the optical power density
required for two-photon excitation.
Confocal principle
Point source
illumination
Beamsplitter
Two-photon
mirror
High-power pulsed
point source
illumination
Beamsplitter
Detector
OCT coherence gate
Broadband
point source
illumination
Beamsplitter
Detector
Detector
Objective
Pinhole
aperture
Objective
Objective
z
y
x
Optical
section
z
y
x
Optical
section
z
y
x
Path length
matched
In confocal imaging, the co-registration of the focus of the source illumination and
detection sensitivity of a pinhole aperture defines the localization region. In two-photon
imaging, the focus of the illumination defines the region of generated signal that is collected by a detector. In OCT, a broadband light source with a short coherence length is
used in a Michaelson interferometer to yield signal only from the region where the path
lengths are matched.
OCT is based on the principle of
low-coherence interferometry. The OCT
optical system is an interferometer that
uses a broadband light source with a
coherence length of a few to a few tens
of microns. It produces a detectable
interference signal when the reference
arm path length closely matches the
path length in the sample. Like confocal
and two-photon microscopy, a two-dimensional image perpendicular to
the optic axis can be built by two-dimensional scanning of the focused
illumination across the sample. This
mode of operation is usually referred to
as optical coherence microscopy.
Alternatively, one can do depth “
scanning” by moving the reference arm mirror (time-domain OCT) or by dispersing
the broadband light onto a detector
array and performing a one-dimensional
Fourier transform (frequency-domain
OCT). A cross-sectional image in the
x-z plane perpendicular to the surface
can be built by depth-scanning combined with one-dimensional beam
scanning. Spatial resolution in OCT
images varies from a few to a few tens
of microns, depending on the numerical
aperture of the beam, which determines
that lateral resolution, and the coherence
length of the source, which determines
the axial or depth resolution.
In tissue, the maximum imaging
depth for all optical biopsy techniques
is ultimately limited by light scatter.
Imaging depth in confocal and nonlinear microscopy is restricted to about
one hundred to a few hundred microns
depending on the tissue type and the
illumination wavelength. OCT can
image somewhat deeper (on the order
of 1-2 mm) due to the higher sensitivity of coherence gating and the use of
infrared illumination wavelengths that
penetrate deeper into tissue. Since many
types of cancer arise in the surface
epithelial layer, optical techniques that
image the tissue surface can be useful
diagnostic tools.
Image contrast
Another fundamental difference
between in vivo imaging of tissue and
conventional biopsy is image contrast.
In conventional biopsy, prepared tissue is stained to produce high contrast
between important subcellular features.
For example, standard hematoxylin
and eosin (H&E) staining produces
dark-blue cell nuclei against light-pink
cytoplasm. The size, distribution and
heterogeneity of nuclei are important
characteristics that pathologists use to
identify disease.
