Spectroscopy/Oncology/Optical Biopsy: Visible resonance Raman improves tumor identification and grading

Optical spectroscopy, including Raman methods, offers many advantages over standard techniques for tissue assessment, including noninvasive, label-free, and real-time operation; time efficiency; and reproducibility. Similarly, resonance Raman (RR) spectroscopy shows advantages over conventional Raman using slight absorption from specific biomolecules in the cells and organelles—such as from proteins, lipids, flavin adenine dinucleotide (FAD), amino acids, reduced nicotinamide adenine dinucleotide (NADH), collagens, elastin, carotenoids, hemeproteins, and mitochondrial cytochromes—these molecules contain fluorophores whose absorption of light leads to resonance enhancement of specific vibrational modes (for example, stretching, bending, and scissoring). The spectra collected from resonance-enhanced molecules can be detected at concentrations <1.0 nM. RR can preferentially target the activities of particular molecular species, making Raman peaks more accessible for identifying changes correlated with various human diseases.

A 2019 book, Neurophotonics and Biomedical Spectroscopy, features exploration of RR for detecting human brain tumor (see Fig. 1),¹ and new results obtained by the authors, along with colleagues in China and the U.S., are the first to use visible resonance Raman (VRR) spectroscopy to target glioma, which is the most common type of central nervous system (CNS) tumor.² Since 2011, researchers from the U.S. and China, respectively, have been collaborating on using VRR to study a variety of human tissues. The work is led by the Alfano group at The City College of New York of the City University of New York, which has done pioneering work since 1987 in applying Raman for optical biopsy. The research work was completed by Zhou’s neurological research team in the Beijing laboratory of Tianjin Raman Medical Technology Center (TRMTC) in China.³ There is collaboration between China and the U.S. to solve the cancer problem of the breast and brain.

Glioma and spectroscopy

Even using a combination of state-of-the-art treatment modalities, mortality and recurrence rates are rising for glioma. Although multiple factors affect outcomes for glioma patients, the ability to accurately delineate tumor boundary for effective resection is critical. So is the grading of gliomas, an active area of research that is expected to improve diagnosis and treatment.

The current gold standard for brain tumor diagnosis—biopsy and histopathology—is an invasive and time-consuming process. Alternative detection methods exist, but each has drawbacks. A review of postoperative MRI results revealed that the accuracy and timeliness limitations restrict full glioma resection to only ~65%. Other studies show that in 75% of cases, there was no ability to clearly distinguish among higher- and lower-grade gliomas.

The high cost of emerging molecular genetic pathology, and the desire to increase accuracy, while reducing time-to-diagnosis, are motivating researchers to explore optical molecular histopathology techniques.

Early studies of RR used ultraviolet (UV) light, usually at <300 nm. VRR, on the other hand, uses a 532 nm laser to enhance the Raman signal intensity from intrinsic molecules by ~100X due to its electronic resonance effect inside the tissue. VRR’s signal-to-noise ratio (SNR) is higher, which makes it possible to directly compare cancerous and normal tissues. There is a large background “wing” that needs to be removed. At the same time, VRR uses less laser power and a shorter integration time to collect signals. Therefore, it is safer for application on live tissues and cells.

Grading and margin assessment

The new article reported the application of VRR to evaluate biomarkers for detecting glioma margins and to investigate correlations between biomarkers and tumor grades. After collecting spectra from various grades of glioma tissues, the team applied machine-learning techniques to classify the specimens, and compared the results to results obtained from traditional histopathology. Analyses based on the measured VRR spectral data revealed that the major molecular biomarkers (including carotenoids, tryptophan, amides, proteins, and lipids) are involved in metabolic processes in human brain, and can form the foundation of an optical pathology method.

Traditionally, gliomas are classified into grades I to IV according to morphology, degree of malignancy, and location of tumor cells. The study revealed that all RR molecular fingerprints occurring in grade I glioma tissue showed a distinct shift in higher grades of glioma tissues (see Fig. 2).

The team established criteria for differentiating tissue types and used these criteria to identify tumor margins, reporting sensitivity at 100% and specificity of 71% compared to chemical reagent-based histopathological analysis.

Next steps

The team continues to make progress in developing and applying the Raman scattering microscopy techniques in studying brain cancer. Specifically, plans are underway to refine the approach by exploring issues observed in the research (for example, 1588 cm^-1 Raman mode) and by identifying real-time correlations between laboratory-derived Raman spectral characteristics and results obtained through examinations at the Beijing laboratory of TRMTC.

The team is also working on multiple instrumentation designs, including a fiber-optic VRR system that measures all molecular spectral biomarkers in real time to detect brain cancer and identify the tumor boundaries in vivo. Moreover, they plan to develop a nonlinear optical microscopy system that will combine VRR and stimulated Raman scattering (SRS) for general optical biopsy applications.

A portable VRR device with a scanner for imaging and with integration time <1 s (see Fig. 3) has already been tested on mice and is currently being used to measure human samples ex vivo. In the animal-model experiments reported at the Frontiers in Optics + Laser Science conference (held Sep. 15-19, 2019, in Washington, DC; paper JW3A.5), the laser power was maintained at 3.5 mW. The final spectral resolution was 8 cm^-1 in the range of interest (200–4000 cm^-1). The 532 nm excitation beam directly excites the surface of the mouse brain (can be tumor or normal brain tissue) for 1 s to obtain a Raman spectrum. A fiber-optic probe with a 200 µm focus spot was used to focus the laser beam and collect the scattered signals from the sample surface—in this case, in vivo mouse. The next step is to apply it on human brain in vivo.

REFERENCES

1. Y. Zhou et al.,“Visible resonance Raman spectroscopy in human brain tissues,” Neurophotonics and Biomedical Spectroscopy, R. R. Alfano and L. Shi (Eds.), 65–106, ISBN 9780323480673. Elsevier, Cambridge, MA (2019); https://doi.org/10.1016/B978-0-323-48067-3.00004-4.

2. Y. Zhou et al., J. Biomed. Opt., 24, 9, 095001 (2019); doi:10.1117/1.jbo.24.9.095001.

3. Y. Zhou et al., J. Biomed. Opt., 17, 11, 116021 (2012).