Rapid, label-free detection of brain tumors with stimulated Raman scattering microscopy

Abstract

Surgery is an essential component in the treatment of brain tumors. However, delineating tumor from normal brain remains a major challenge. We describe the use of stimulated Raman scattering (SRS) microscopy for differentiating healthy human and mouse brain tissue from tumor-infiltrated brain based on histoarchitectural and biochemical differences. Unlike traditional histopathology, SRS is a label-free technique that can be rapidly performed in situ. SRS microscopy was able to differentiate tumor from nonneoplastic tissue in an infiltrative human glioblastoma xenograft mouse model based on their different Raman spectra. We further demonstrated a correlation between SRS and hematoxylin and eosin microscopy for detection of glioma infiltration (κ = 0.98). Finally, we applied SRS microscopy in vivo in mice during surgery to reveal tumor margins that were undetectable under standard operative conditions. By providing rapid intraoperative assessment of brain tissue, SRS microscopy may ultimately improve the safety and accuracy of surgeries where tumor boundaries are visually indistinct.

Figure 1. Two-color SRS microscopy

(A) Energy diagram of the SRS process, where pump and Stokes photons excite the ground state (ν=0) molecules to their vibrational excited state (ν=1), resulting in the reduction of pump intensity–stimulated Raman loss (SRL) and the increase in Stokes intensity–stimulated Raman gain (SRG). (B) Raman spectra from frozen sections of a mouse brain with human GBM xenografts show white matter, cortex, and tumor. The marked frequencies at 2845 cm⁻¹ and 2930 cm⁻¹ were chosen for two-color SRS imaging. (C) Experimental setup of epi-SRS microscopy. Stokes beam was modulated at high frequency (10 MHz), and the weak SRL signal was demodulated by a lock-in amplifier. Epi-detection scheme was used for in vivo brain imaging and ex vivo imaging on fresh tissues. CS, coverslip; DC, dichroic mirror; EOM, electro-optical modulator; FI, optical filter; PD, photodiode; SL, saline. (D) Neurons in gray matter were imaged at 2845 cm⁻¹ (left) and 2930 cm⁻¹ (middle). A linear combination of the two raw images was used to compute the distributions of lipid (green) and protein (blue), shown in a composite image (right).

Figure 2. SRS and H&E images of frozen normal mouse brain sections

Images are representative of 6 mice. Images were taken with SRS microscopy, and then stained with H&E for comparison. Lipids have been false-colored green; proteins in blue. (A) A full coronal section of normal mouse brain. (B) SRS microscopy demonstrates major structural features of the normal brain, such as the hippocampus shown here. (C) SRS microscopy differentiates regions of the brain based primarily on cellularity and the relative presence of lipids and proteins. Cortex (dagger), white matter (asterisk), and the CA1 region of the hippocampus (arrow) are readily identified. (D) The gray/white junction is evident owing to the differences in lipid concentration between cortical and subcortical tissue, with cell-to-cell correlation between SRS and H&E images (circles).

Figure 3. SRS and H&E images of frozen human GBM xenografts

Images are representative of 6 mice. Images were taken with SRS microscopy, and then stained with H&E for comparison. Lipids have been assigned to the green channel and proteins to the blue channel. (A) A thin (10 µm), full section of snap-frozen brain from implanted human GBM xenograft in mice. (B) High-magnification view of normal to minimally hypercellular cortex. (C) Infiltrating glioma with normal white matter bundles (asterisk), tumor-infiltrated bundles (arrow), and dense tumor cells (arrowhead). (D) High-density glioma. FOVs similar to these were used to populate the Web-based survey to quantitatively compare SRS and H&E microscopy.

Figure 4. Epi-SRS images of fresh brain slices from normal mouse and human GBM xenograft (BT112) mouse model

Images are representative of 10 normal mice (A to H) and 6 BT112 mice (I to M). Lipids have been assigned to the green channel and proteins to the blue channel. Fresh 2-mm thick coronal section of normal mouse brain. Structural features of the full section from a normal mouse (A), cortex (B), hippocampus (C), corpus callosum (D), choroid plexus (E), hypothalamic nuclei (F), habenular nucleus (G), and caudatoputamen (H) demonstrate the expected histoloarchitectural patterns. In contrast, fresh 2-mm-thick coronal brain section of a BT112 human glioblastoma xenograft-bearing mouse reveals normal, green appearing brain parenchyma surrounding blue hypercellular tumor (I). High magnification of the tumor core reveals individual tumor cells (J). The tumor-gray matter interface (white dashed line) demonstrates an invasive pattern of tumor growth (K). The tumor-white matter interface, demonstrates the ability of tumor cells (blue) to traverse and separate white matter bundles (green) (L). A line profile of S₂₉₃₀/S₂₈₄₅ across the gray matter/tumor interface in (K, red dot-dashed line) shows higher S₂₉₃₀/S₂₈₄₅ with increasing tumor density (M).

Figure 5. In vivo SRS microscopy images of human GBM xenografts

Images are representative of 6 mice. SRS imaging was carried out via acute cranial window preparation in mice 24 days post-implantation of human GBM xenografts. (A) Bright field microscopy appears grossly normal, whereas SRS microscopy within the same FOV demonstrates distinctions between tumor-infiltrated areas and non-infiltrated brain (normal), with a normal brain/tumor interface (dashed line). (B to D) High-magnification views within the tumor (B), at the tumor/brain interface (C), and within normal brain (D).

Figure 6. SRS imaging during simulated tumor resection on mouse brain

En face, epi-SRS images were obtained in vivo during various stages of a simulated tumor removal. The cartoons on the right show the depth of imaging. In a tumor located beneath the cortical surface, there is no obvious abnormality in SRS (left) or brightfield images (middle) when imaging the cortical surface (A). After a portion of the cortex has been removed, the tumor is revealed. Blood was present on the dissected surface, but did not adversely affect the distinction from tumor-infiltrated regions from non-infiltrated regions (B). As dissection was carried deep past the tumor, the normal appearance of white matter and cortex was again visible (C).

Figure 7. SRS and H&E microscopy of freshly excised tissue from a human brain tumor

SRS images were obtain from freshly excised human glioblastoma specimen and compared to similar regions in H/E stained tissue from the same specimen. The hypercellularity of viable tumor (A) contrasts with normocellular regions of adjacent brain with with minimal tumor infiltration (B). Higher magnification images of the different regions in the specimen demonstrate key diagnostic features of glioblastoma including cellular pleomorphism (C), pseudopallisading necrosis, where densely cellular regions (arrow) border bland, acellular regions of necrosis (asterisk) (D); and microvascular proliferation (E).