Biomedical Engineering Reference
In-Depth Information
collagen localized around small tumors during early disease; (ii) collagen fibers that are parallel to the
tumor boundary for
in situ
carcinoma; and (iii) collagen fibers that are normal to the tumor boundary
for invasive disease. This approach relies on alignment and the results may be tissue specific, hence the
generalization to other cancers requires further investigation.
6.1.1.3 Signal Processing Approaches to SHG Microscopy of Musculo-Skeletal Disorders
There is considerable interest in signal processing methods, for example, Fourier transforms, wavelet
transforms, and texture analysis that can analyze and classify whole images without examining indi-
vidual features. In healthy muscle fibers, bands of sarcomeres composed of actomyosin complexes are
straight and evenly spaced (~2-3 microns), with each band lying nominally orthogonal to the axis of
contraction. Damaged cells display a range of visible deviations from this norm. To measure changes
that accompany muscle diseases, Plotnikov et al. [22] applied the Helmholtz equation for wave number
to calculate the local striation spacing and angle of orientation with respect to the long axis of the myo-
fiber (90° for ideal case). They tested this approach by comparing controls to three models of muscular
disorders of varying severity: disuse-induced atrophy, mild and severe hereditary muscular dystrophy,
and sarcopenia of aging. Analyzing these images for all the models, we found a consistently negative
correlation between severity of the disorder and the mean sarcomere length. While quantitative, the
scheme relies on the periodicity and alignment of sarcomeres. However, this regularity does not exist in
most tissues and more general approaches are still needed.
6.1.1.4 need for a More General Approach
The analysis approaches described earlier have all demonstrated potential for discriminating normal and
diseased tissues. However, they rely on fiber morphologies and may not be applicable for all cases, espe-
cially if the morphologies are too complex or irregular to quantify. As an alternative, our lab developed a
general approach to quantify 3D (up to several hundred microns) SHG imaging data. In a tissue imaging
experiment, the measured SHG signal is composed of a convolution of the initially emitted SHG photons
as well as the subsequent scattering of these photons at λ
SHG
. The initial SHG emission from tissues has
a distribution of emitted forward and backward components, whose ratio we denote
F
SHG
/
B
SHG
, which
depends on the regularity of the fibril/fiber assembly. Additionally, more ordered tissues will give rise
to brighter SHG due to higher photon conversion efficiency. Collectively, we refer to the SHG emitted
directionality and the emitted intensity as the creation attributes. Following generation of SHG in tis-
sue, the generated photons will propagate based on the scattering coefficient and scattering anisotropy at
λ
SHG
. The scattering coefficient μ
s
is a measure of density, where it is the inverse of the average distance a
photon will propagate before undergoing a scattering collision and changing direction. For most tissues,
these scattering lengths are typically ~20-50 microns in the visible/near infrared region of the spectrum.
The scattering anisotropy,
g
, is related to the directionality of the scattering, and varies from 0 to 1,
where higher values correspond to greater organization. While scattering limits the achievable depth in
any microscopy experiment, there is useful information in the scattering as well. Like the SHG creation
attributes, the scattering properties also depend on tissue structure, and can differ between normal and
diseased states.
The description of tissue scattering in terms of bulk optical properties is well developed and, as will
be explained in a later section, we incorporate previous treatments into our approach. However, no com-
plete, rigorous description of the SHG creation in tissues has been previously given. This is an essential
step for this imaging modality to become a useful clinical diagnostic tool for monitoring disease sever-
ity and progression. Our overall premise is that different fibril/fiber size and packing will be different in
disease states resulting in SHG properties. Examination of the collective findings of the literature has not
yielded a unified relationship between SHG directionality, fibril morphology, and fibril size, suggesting
that fibril size considerations alone are insufficient for complete SHG image interpretation. To solve this
problem, we presented a new heuristic model based on phase-matching consideration, which provides
a mathematical framework leading to the necessary insight to enable a thorough understanding of the
