Figure 1.
Imaging of experimental human glioblastoma and brain metastases in a mouse model.
A: CARS image of a whole mouse brain section with an experimental human U87MG glioblastoma; white matter tracts appear brighter in the CARS image, while the tumor (arrow) is darker that the normal brain tissue. # hippocamus, + corpus callosum, **mesencephalon, *cerebellum B: Magnified CARS image of the area indicated in A; the difference in the CARS signal intensity between the neoplastic tissue and the surrounding white matter enables discerning the tumor border. C: CARS image of a breast cancer metastasis (induced by MCF-7 cells) in a mouse brain. D: CARS image of a metastasis of melanoma (induced by A375 cells) in the hippocampal region of a mouse brain. E–F: H&E stainings of consecutive sections corresponding to A–D. The arrowheads indicate nuclei-rich layers in gray matter of the hippocampus.
Figure 2.
Imaging of the infiltrating tumor margin.
A: CARS image of a human U87MG glioblastoma in a mouse brain. B: CARS image of a separate small glioblastoma island in a mouse brain. C/D: Anti-Ki67 immunohistochemical staining corresponding to A/B. In both examples the very same section was used for CARS imaging and for staining.
Figure 3.
Quantification of the CARS signal intensity of infiltrative tumor areas.
A: Unprocessed CARS image of a human U87MG glioblastoma in a mouse brain, displaying a solid tumor and a large infiltrative region. B: CARS signal intensity along the area indicated in A. The CARS signal intensity range indicative of normal tissue is underlined in green, the intensity range indicative of infiltrative areas in yellow and the intensity range indicative of tumor in red. C: H&E staining corresponding to A.
Figure 4.
Quantification of the CARS signal intensity and lipid-related Raman band intensity.
A: Dot plot showing the CARS signal intensity in normal gray matter vs. the intensity of the CARS signal in the neoplastic tissue for each sample. B: CARS signal intensities of tumors normalized to the respective intensities in gray matter. C: Average Raman spectra of normal gray matter, glioblastoma, melanoma, and breast cancer metastases. D: Intensity of the Raman band generated by symmetric stretching of the C-H bonds in CH2 groups at 2850 cm−1, calculated as integral in the range (2850±15) cm−1. Values were normalized to values of gray matter. B/D: Bars represent mean ± SD, GBM n = 8; melanoma metastases n = 4; breast cancer metastases n = 4; * significant difference vs. gray matter: P<0.05; *** significant difference vs. gray matter or as indicated: P<0.001.
Figure 5.
Tumor-induced changes influence the CARS signal intensity.
A/B: Number of nuclei/mm2 representing cellular density (A) and size of single nuclei (B) within normal gray matter and different experimental tumors in a mouse brain. Bars represent mean ± SD, GBM n = 8; melanoma metastases n = 4; breast cancer metastases n = 4; *** significant difference vs. normal gray matter: P<0.001. C: Dot plot showing the total area occupied by cell nuclei vs. the normalized CARS signal intensity of the respective tumor. D: CARS image of an experimental GBM induced in a mouse brain. E: Anti-CD31 staining of a consecutive section of the one shown in D. F: Overlay of CARS (red) and anti-CD31 (in false color: green). Coarse blood vessels (white arrowheads) and fine blood vessels (gray arrowheads) are detected in the CARS image. Very fine (normal) blood vessels do not cause any alterations of the CARS signal (black arrowheads).
Figure 6.
Quantification of the CARS signal in human GBM.
A: Unprocessed CARS image of a cryosection of a human GBM specimen obtained during routine surgery. The CARS image displays the margin of a solid tumor and an infiltrative region. B: CARS signal intensity along the area indicated in panel A. The range of CARS signal intensity of normal tissue is underlined in green, of infiltrative areas in yellow, and of tumor in red, respectively. C: Dot plot showing the CARS signal intensity in normal gray matter vs. the intensity of the CARS signal in human GBM for each sample.