Biomedical Engineering Reference
In-Depth Information
TABLE 15.1
Comparison of Different Techniques Used for Fibrosis Quantization
Masson's
Trichrome
Picrosirius
Red
Picrosirius
↻polarization
Immunochemical
Labeling
Confocal
Reflectance
OCT
SHG
⊖⊖⊖
⊖
⊕⊕
⊕⊕⊕
⊖⊖
⊖⊖
⊕⊕⊕
Specificity to
fibrillar collagen
3D capability
No
No
No
No
⊕⊕
⊕⊕
⊕⊕⊕
Multimodality
⊖
⊖
⊖⊖
⊕
⊕
⊖⊖
⊕⊕⊕
Reproducibility
⊖
⊖
⊖
⊖⊖
⊕⊕⊕
⊕⊕⊕ ⊕⊕⊕
Cost-Complexity
⊕⊕⊕
⊕⊕⊕
⊕
⊕
⊖⊖
⊕
⊖⊖⊖
1979, Whittaker et al. 1994). Sirius Red dye indeed aligns along the collagen molecules and enhances
the fibrillar collagen birefringence that is detected in polarized light microscopy. The specific advantage
of using circular light is the same as in SHG microscopy: it enables homogeneous imaging of fibrils with
different orientations in the focal plane. In practice, this technique is an excellent choice when quanti-
fying thin sections from human biopsies as a trade-off between cost and efficiency. SHG microscopy is
however a better choice when imaging the 3D distribution of fibrosis in thick tissues and comparing it
to the localization of other components of interest. For this purpose, SHG microscopy takes advantage
of its 3D capability and multimodality. Confocal reflectance microscopy and optical coherence tomog-
raphy (OCT) offer the same advantages but lack specificity to fibrillar collagen.
The sensitivity of SHG microscopy and of other techniques has not been fully characterized since it
is a complex task. SHG yields the unique advantage of highly contrasted images because of its coherent
nonlinear nature (see Section 2.2). It therefore favorably compares to confocal reflectance microscopy or
OCT. A rigorous characterization of SHG sensitivity would however require determining the minimal
diameter of fibrils that are detected in SHG microscopy. It has not been reported yet to the best of my
knowledge. Note that the sensitivity of histological techniques is also poorly characterized.
All these considerations are summed up in Table 15.1. A more rigorous and complete comparison
of SHG with other techniques would require using the same standard of image processing in order to
distinguish the advantages of the technique itself and the ones due to image processing. Since image
processing of histological sections is usually quite basic, the comparison to SHG is somewhat biased
and histology should benefit from advanced segmentation and quantization techniques. For instance,
automatic quantization of Masson's trichrome-stained renal biopsies based on a colorimetric segmenta-
tion algorithm has been reported recently to significantly improve the diagnosis of long-term allograft
disease (Servais et al. 2007). In liver pathology, this approach also allowed to detect beneficial effect of
antiviral treatments in chronic viral hepatitis that was otherwise not detected with semiquantitative
standard evaluation of liver fibrosis (Goodman et al. 2009). I however anticipate that SHG imaging will
prove to be more sensitive and reproducible than all other techniques.
15.5.2 Perspectives, Possible improvements
15.5.2.1 true Quantization
The main limitation of fibrosis quantization by means of SHG microscopy is its inability to retrieve the
quantity and 3D distribution of collagen molecules within the focal volume. Actual SHG imaging of
fibrotic pathologies quantifies the extent of fibrosis, but it does not make use of the SHG signal intensity.
The SHG signal is obtained as the coherent summation of the second-harmonic fields radiated by all the
harmonophores within the focal volume. Truly quantitative measurements would require to know the sec-
ond-order response of one single collagen molecule, or better of the harmonophores (the peptide bonds)
within this molecule, and to solve the inverse problem of the second-harmonic fields coherent summation
over the 3D distribution of collagen molecules within the fibrillar macro-organization.
