Biomedical Engineering Reference
In-Depth Information
is quite sensitive to many parameters affecting the yield of the immunochemical reaction (accessibility
of the epitope, tissue preservation, section thickness, etc.). Generally speaking, any staining procedure,
including histological staining, strongly limits the reproducibility of fibrosis scoring because of issues
such as dye preservation, section thickness, and so on. In that respect, SHG microscopy is the only spe-
cific and reproducible technique for fibrosis scoring because it applies to unstained tissues.
The second limitation regards the image analysis. To improve the reproducibility of fibrosis scoring
between different clinical or research centers, image analysis must be automated and provide quantita-
tive scores. There is a lot of activity in that field to develop automated methods based on image analysis
to quantify fibrous tissue amount (Pilette et al. 1998, Grimm et al. 2003, Pape et al. 2003, Sund et al.
2004, Friedenberg et al. 2005, Matalka et al. 2006, Goodman et al. 2007, Servais et al. 2007, 2009).
Morphometric analysis provides continuous indexes of the fibrosis extent (the ratio of fibrosis area to
the total area of the imaged section) instead of the conventional classification in a few grades (ci0 to ci3
in the Banff classification of kidney allograft pathology, F0−F4 in Metavir score of liver hepatitis C).
The main difficulty however stems from the variability in the colorations that requires advanced color
segmentation. On the contrary, SHG microscopy provides gray-level images of the fibrosis that can be
analyzed in a more direct way. Basic thresholding of the SHG image and calculation of the mask area
are sufficient to obtain a score of the fibrosis extent. More advanced image processing is however useful
to increase the reproducibility and sensitivity of SHG scoring, as presented below. Note however that
neither SHG microscopy nor conventional techniques provide properly quantitative fibrosis scores. All
these techniques assess the extent of fibrosis but do not quantify the number of collagen molecules
accumulating in the tissue. Indeed, the SHG signal depends in a complex way on the distribution and
density of collagen in the focal volume as explained in the former sections. Proper quantization of col-
lagen accumulation is a challenging issue that requires complementary experimental and theoretical
work (see perspectives below). Consequently, phenomenological approaches have been preferred yet and
SHG in practice quantifies the extent of the fibrillar collagen network within the tissue.
15.4.2 experimental Setup and Protocols
Fibrosis quantitative imaging was first proposed using custom-built multiphoton microscopes. It can
also be implemented in recent commercial setups with sensitive nondescanned detection channels if
fibrosis detection is optimized as follows:
• The laser excitation has to be circularly polarized to obtain a homogeneous SHG image from fibrils
with various orientations within the focal plane. SHG is indeed described by a third-rank tensor
and is sensitive to the orientation of the excitation electric field relative to the fibrils orientation.
The way to achieve a circular polarization is to put a quarter waveplate in the laser excitation path.
It is better to use an achromatic waveplate and to put it at the back pupil of the objective lens. Due
to the ellipticity introduced by the laser scanning components, the achieved polarization will not
be perfectly circular, unless ellipticity is corrected by another quarter waveplate (Gusachenko
et al. 2010). However, it is sufficient to mitigate orientation effects and to image all the fibrils
within the focal plane with similar efficiency. Note that advanced polarization shaping of the laser
excitation may be used to image out-of-plane fibrils (Yew and Sheppard 2007, Yoshiki et al. 2007).
• SHG signal from thick fibrotic tissues can be detected either in the forward direction or in the
backward direction. Backward detection takes advantage of scattering processes that partly redi-
rect forward-directed SHG signal in the backward direction. It is more efficient using an objective
with a large field of view to better collect scattered signals (Beaurepaire and Mertz 2002). SHG
images of thin sections are better recorded in the forward direction because SHG signals are more
important in the forward direction and scattering is negligible in thin samples.
• The choice of the objective is an important issue since it must combine a large field of view to
visualize significant regions of the fibrotic tissue and a good resolution to detect thin collagen
