Biomedical Engineering Reference
In-Depth Information
The ability to probe ECM structure in diverse tissues gives the SHG imaging modality a great poten-
tial as a clinical diagnostic tool. In terms of human health, the greatest impact may be in early cancer
detection. SHG has already shown early promise in imaging cancer since malignant tumors often have
abnormal assembly of collagen relative to normal tissue [4,13-17]. This has now been shown in both ani-
mal models as well as human tissues
ex vivo
and
in vivo
. Besides simple visualization of fiber morphol-
ogy, an additional enabling property for diagnostic imaging arises from the coherent nature of the SHG
process. This is manifested in the initial directionality of the emission, where the morphology observed
in the forward and backward channels is reflective of the fibril size distribution as well as the order of the
packing. This is of particular importance as the fibril size and distribution may be different in healthy
and diseased tissues, and we have shown this to be the case for the oim murine model for OI [12,18] and
more recently in ovarian cancer [14].
The results from our lab and those of other labs now suggest that SHG has the potential to be devel-
oped into a clinical tool to analyze
ex vivo
biopsies or to perform
in vivo
imaging through endoscopes.
A large remaining challenge is how to quantify and standardize 3D SHG image data for these diagnostic
purposes. In this chapter, we describe our efforts to provide a general approach for measuring and mod-
eling 3D SHG data to differentiate normal and diseased tissues based on SHG creation physics as well as
photon propagation. We begin by describing some other analysis methods and discuss their limitations.
Next, we describe experimental and modeling methods, present a theoretical treatment of SHG process
in tissues, and then show examples of the general approach for OI and ovarian cancer, and finally con-
clude with some perspectives.
6.1.1 Limitations of existing SHG Analysis techniques
6.1.1.1 comparison with Histological Analysis
A primary motivation for pursuing SHG for biomedical imaging lies in the ability to provide more
quantitative/less subjective analysis than possible by classical histology. SHG microscopy has several
advantages over standard histological scoring procedures for diagnostic imaging. First, SHG microscopy
acquires 3D image sets through tissues of several hundred microns of thickness and can obtain more data
than possible by histologic sections. In the latter, tissues are fixed, sliced in cross section into thin slides
(~5-10 microns in thickness), and stained. The process can lead to artifacts. Perhaps more importantly,
the interpretation depends highly on the skill of the pathologist. Comparisons between histology and
SHG have been reported for several tissues, mostly for the purposes of visual assessment. More quantita-
tive analyses of the collagen from histological sections have been reported for fibrosis. The first demon-
stration of SHG microscopy for scoring fibroses was done in a mouse model by Schanne-Klein [19]. For
quantification of the collagen changes, they used a thresholding process for image segmentation to iden-
tify individual fibers. Through this process, they were able to successfully discriminate normal versus
fibrotic tissues. A somewhat more sophisticated segmentation approach was taken by Yu and coworkers
for imaging liver fibrosis in a rat model [20]. They used Otsu segmentation to score the amount of result-
ing collagen. The results of their automated approach were compared to pathology analysis and revealed
a good correlation between these approaches, especially in areas of low collagen coverage. These results
are promising but the specificity and generality of the analysis requires further study.
6.1.1.2 Quantification of Fiber Alignment in cancer
It has been suggested that changes in collagen alignment in the ECM during cancer progression can
be a useful metric. This is because there is an increase in collagen deposition, or desmoplasia, in many
epithelial cancers. To determine if changes in collagen can be an early diagnostic of breast cancer, and,
further, if SHG is sensitive to these alterations, Keely and coworkers measured the alignment of collagen
fibers in murine tumor models of over a range of disease progression. In these efforts, they characterized
three “tumor-associated collagen signatures (TACS),” which are reproducible during defined stages of
tumor progression [16,21]. These signatures (1, 2, and 3) are characterized by (i) the presence of dense
