Raman Spectroscopy and its Clinical Use
Potential and Limits
- Fig. 1: Raman spectra of the main biological constituents of bacteria: water, protein, nucleonic acids (DNA), carbohydrates and fats. In the illustrated examples of Raman spectra of various strains of staphylococcus, the various bands of the individual constituents can be seen.
- Fig. 2: Microfluidic chips made of quartz, for Raman-activated sell sorting.
- Fig. 3: Comparison of TPEF, CARS and Raman microscopy of an unstained microtome section of a brain tumour with the optical microscopy image of the sample, which was subsequently stained with haematoxylin and eosin.
- Fig. 4: Raman endoscopy examination of the arteries of a rabbit.
Raman spectroscopy is a promising method for the solution of clinically relevant challenges in the field of pathogen diagnostics, oncology and cardiovascular disorders thanks to its molecular sensitivity. Corresponding approaches will be presented in this article.
In the 21st century, due to an ageing population, we are facing the challenge of ensuring an affordable and sustainable health care system. To meet this challenge, we require the development of new methods and equipment, with which illnesses can preferably be detected and counteracted before they become manifest. Here, photonics plays a key role. An especially promising and versatile photonic method is Raman spectroscopy and its variants. As in general with all light-based methods, Raman spectroscopy enables contact-free measurement, which however, in contrast to e.g. fluorescence spectroscopy, also functions without exogenous labels. Because of this, Raman spectroscopy has the potential to enable diagnostics close to the patient, especially as it is not only a comparatively fast, but is also a very precise method. Especially in the field of imaging methods, the high level of specificity and the low invasiveness must be emphasized. In addition, Raman spectroscopy has the advantages of high spatial resolution, it does not require extensive sample preparation and offers the possibility of working in an aqueous environment.
The following article presents some selected examples of especially promising application possibilities of Raman spectroscopy in the medical and clinical field.
Traditional pathogen diagnostics is based on creating and analyzing a pathogen culture, which can take up to a week and requires experienced specialists. In some cases of infections, e.g. a sepsis, the survival rate reduces dramatically with each hour that passes before a targeted treatment is possible. The ideal case would be that an infection could be attributed to a pathogen within a few hours. Due to the fact that each species of bacteria has its own individual Raman signature, Raman spectroscopy can be used for the identification of bacteria, whereby the spectrum of an individual bacterium is sufficient for its identification.
However, the differences between the spectra of various species are often subtle and assignment according to visible appearance is not possible. Ultimately, the spectrum of the bacterium is the sum of the spectral signatures of all of the substances which it contains, such as water, proteins, fats, nucleic acids, carbohydrates, etc. (fig. 1).
Therefore there are subtle differences even between the spectra of different bacteria of the same species, due to e. g. age, state of nutrition and various environmental influences. The solution to this problem consists of the application of chemometric methods to the Raman spectra of the bacteria. Here, to put it simply, a spectrum is broken up into especially prominent areas and these are compared mathematically with the equivalent areas of bacterium spectra, which are collected in an extensive database. On average, not only can the species of almost 99 % of the bacteria be identified, but also even the strain of the bacteria within a species can be determined with an average value of more than 92% . A corresponding solution is already commercially available for the determination of contamination in clean rooms or in air conditioning systems (Biopartikelexplorer, Rapid Particle Systems). This uses fluorescence spectroscopy to differentiate between inanimate particles and bacteria. The bacteria are then identified by means of Raman spectroscopy. However, use in hospitals makes it necessary to identify bacteria in more complex media such as saliva, urine or even blood. For this, the bacteria must be separated from these media, as otherwise the medium would make identification difficult if not impossible. At present, microfluidic chips are being developed for this separation stage, which for example utilize dielectrophoresis in order to capture bacteria and make them accessible for measurement . However, the latter method can also be used to measure bacteria directly in solution and therefore, in addition to identification, it also enables e.g. the possibility of making statements with regard to susceptibility or resistance to antibiotics.
Tumor cells may detach from a carcinogenic tissue, enter the bloodstream and finally cause metastases. These individual tumor cells are relatively easily accessible and are very valuable for diagnosis. The diagnostic principle essentially corresponds to that of optical flow sorting. The blood is passed through a microfluidic chip. Individual cells in the blood flow are trapped by optical traps, where they are examined and classified using Raman spectroscopy and sorted for further use. In contrast to optical flow sorting, with the aid of Raman spectroscopy it is now possible to make a considerably more accurate diagnosis of the individual cell. However, the price for this advantage is a considerably lower throughput (5 - 6 cells / minute), which however could be considerably improved in the future by equipment development. Figure 2 shows a microfluidic chip made from quartz for Raman-activated cell sorting . Individual cells are trapped in an optical trap by means of a laser. After Raman-based identification, the cell can be passed on for sorting. The unprocessed Raman spectra are influenced by the spectral characteristics of filters, the capture laser and the substrate material. Successful classification therefore requires suppression of these characteristics, so that the spectral fingerprint of white blood cells (green) and tumor cells (orange, brown, blue) is visible.
Particularly, in the case of brain tumors, radical removal of the tumor is not advisable. Instead, as little healthy tissue as possible should be removed. However, the boundaries between the tumor and healthy tissue are often diffuse. Up to now, tissue removed in a biopsy was assessed by a pathologist, and the decision as to whether or not the tumor should be removed was made on the basis of this assessment. It would therefore be desirable to have a surgical microscope, which would directly depict the boundaries of the tumor. Such a microscope could be realized by a combination of several imaging techniques. For this, several, preferably label-free techniques are combined in order to increase the contrast. The combination of Raman with coherent Anti-Stokes-Raman-Scattering (CARS) and two-photon fluorescence excitation (TPEF) with the use of endogenous markers and the "Second Harmonic Generation" (SHG) is considered to be especially promising . While with Raman scattering, all (Raman-active) vibrational modes are excited simultaneously, with CARS, a selected vibration is isolated and coherently excited due to the interaction of three different pulses of light, whereby a fourth, spatially oriented pulse of light is generated. The advantage over Raman imaging lies in the significantly reduced time for the recording of an image (up to a factor of 104) due to a greatly increased scattering cross-section. In contrast to Raman and CARS, SHG and TPEF emphasize morphological details. SHG is especially sensitive to ordered, non-centro-symmetric structures, e. g. collagen, where in contrast TPEF highlights endogenous fluorescent molecules such as NAD(P)H, flavin, elastin etc. Figure 3 compares TPEF, CARS and Raman microscopy of an unstained microtome section of a brain tumor with the optical microscopy image of the sample, which was subsequently stained with haematoxylin and eosin. In particular the cell nuclei, which are resolved by all methods, are of interest for histo-pathological evaluation. With the aid of combined morphological and functional information from the multimodal approach, there is a good chance that photonic tools can be developed, which on the one hand are able to detect and classify tumors at an early stage, and on the other hand enable the localisation of tumor boundaries e. g. during surgery to remove the tumor.
Diagnostics of Cardiovascular Disorders
Here, the focus is on future in vivo applications of Raman spectroscopy for the endoscopic examination of plaques in arteries. For the assessment of whether a deposit is dangerous, e. g. if it can detach from the wall of the blood vessel and cause blockages, an assessment of the morphology is not sufficient. As the Raman spectra of calcium phosphate, connective tissue, triglycerides and cholesterol are easily differentiated, in principle, Raman spectroscopy enables the determination of the composition of plaque, and therefore the danger of the deposits. Initial experiments with rabbits confirm the value of this approach. As illustrated in figure 4, a probe with a diameter of 1 mm with a central excitation fibre and 12 detection fibres was used for ex vivo measurements . The Raman spectra were recorded with this Raman probe under in vivo conditions. The signals from the plaque deposits differ in the intensity and spectral contributions of lipids in contrast to the signals from the walls of the artery with collagen bands and from blood, with bands from red blood cells. Future developments focus on the combination of Raman with optical coherence tomography and/or ultrasound, in order to combine molecule-specific information with the morphological information. In addition, miniaturisation is aimed for, in order to be able to investigate arteries with smaller diameters.
 Harz M. et al.: Cytometry A 75A, 104-113 (2009)
 Schröder U.-Ch. et al.: Infection 41 (Suppl.1), P036 (2013)
 Dochow S. et al.: Anal Bioanal Chem. 405, 2743-2746 (2013)
 Meyer T. et al.: J. Biomed. Optics 16, 021113 (2011)
 Matthäus C. et al.: Anal. Chem. 84, 7845-7851 (2012)