Katsumasa Fujita

	Although the concept of Raman imaging was proposed many years ago, the evolution in ultrafast laser technologies, high-sensitivity detectors, and ultrafine optical components provided ways to boost the weak Raman signal so that we can fully utilize the power of Raman spectroscopy in microscopy analysis. 

	Raman imaging has already been recognized as a new imaging modality in research fields that can compensate for the biological information lost in the observation using conventional techniques. The capability of detecting molecular vibrations in Raman microscopy has enabled label-free imaging of sample morphology in three-dimensional (3D) and comprehensive analysis of intrinsic molecules. These characteristics allow the investigation of untreated biological samples under physiological conditions, leading toward the realization of anticipated biomedical applications, such as intraoperative rapid diagnosis and tissue or cell qualification for regenerative medicine and drug discoveries. Raman microscopy provides unexplored spectroscopic insights into biological samples that attract scientists to invent new tools for biomedical studies.

	Raman microscopy has brought new strategies for labeling molecules. Functional groups that exhibit Raman peaks in the so-called 

 can be used to label small molecules and observe them by Raman microscopy. Small molecules, such as nucleic acids, lipids, drugs, and so on, were not observable using fluorescent probes, which are typically larger than the targets. Furthermore, the narrow emission spectrum of a Raman tag has enabled super multiplex imaging of intracellular molecules or structures. The Raman tag technique expanded the toolkit for imaging molecules in biological samples.

	Currently, we have several different tools for Raman imaging, but still need technology developments, especially improvements in sensitivity. Although there are many techniques for boosting the signal, detecting low concentration molecules (<100 µM) is difficult. Using surface-enhanced Raman scattering (SERS) would be a promising route to enhance sensitivity. However, the fluctuation of signal intensity and Raman peak positions hinders the expansion of applications. This issue might be solved by using the power of computation that is continuously being developed. Increased computational power also benefits the interpretation of complex Raman spectra from biological samples, and, for this purpose, it is necessary to build a reliable database of Raman spectra of biological samples in various layers, from molecules to tissues. The recent development of quantum cascade lasers has made it easy to perform infrared absorption imaging, which might be a solution for detecting low-concentration molecules using vibrational contrasts.

	Another important issue in Raman microscopy is usability. Many studies using Raman spectroscopy or microscopy require professionals in spectroscopy for the operation of instruments and interpretation of spectra. The cost of Raman microscopes is relatively high (especially to buy or even build coherent Raman microscopes). Technology developments for low-cost Raman microscopes and improving usability, including development of spectral databases and spectrum analysis software, are important to enable biomedical Raman imaging to leap forward to become a standard tool in biomedical science and industry.

 is a professor in the Department of Applied Physics at Osaka University, in Osaka, Japan. Direct correspondence to 

Articles

In the past decades, we have witnessed the evolution of imaging technologies based on vibrational spectroscopy. In particular, the technical developments in Raman, coherent anti-Stokes Raman spectroscopy (CARS), and stimulated Raman scattering (SRS) microscopy allow researchers to gain new insights in biological, medical, and pharmaceutical studies.

A Further Leap of Biomedical Raman Imaging

) was held in Osaka, Japan,  November 24–26, 2019. The symposium was aimed to discuss recent advances and technological developments in Raman imaging and its application in biological and medical research. In total, 129 participants gathered from Asia, Europe, and the United States and introduced their recent research progress and shared the perspectives on the research field. The technical program was accompanied by an exhibition that included 14 companies providing instruments and devices for Raman microscopy and spectroscopy and their application.

	Raman spectroscopy has  been utilized as a powerful technique for material analysis, but due to the low scattering efficiency of the technique, its contribution to the biomedical field has been limited. Recent technical development, especially those focused on imaging, pave a way to applications in biomedical fields, where morphological information of samples provides rich information. The imaging capability makes it possible to combine morphological and spectroscopical information to explore biological functions and causes of disease.

	Biomedical Raman Imaging 2019, held in Osaka, Japan, November 24–26, 2019, was kicked off with a keynote talk by Ji-Xin Cheng from Boston University. He introduced the history of coherent Raman microscopy and his achievements in fast and analytical imaging of biological samples. In coherent Raman microscopy, two different wavelengths of laser are utilized to introduce the Raman effect efficiently through a stimulated transition of molecules to the vibrationally excited state. The enhancement of the signal can reach six orders of magnitude, which allows us to rapidly scan the sample by a focused laser to form a Raman image. Coherent Raman microscopy drastically boosts the light-to-signal conversion efficiency by exciting vibrational modes of interest selectively.

	On the other hand, techniques using spontaneous Raman scattering take different strategies. Because spontaneous Raman spectroscopy is performed with a simple illumination–detection scheme, the total signal can be easily boosted by parallel detection of Raman spectra. In Raman imaging, this advantage is utilized in the spatial multiplicity of Raman detection. Katsumasa Fujita introduced the parallel detection of Raman spectra from a sample irradiated by a line-shaped focus and improved the image acquisition speed for detecting the fingerprint and high-wavenumber region. Ioan Notingher of the University of Nottingham described a technique that uses a digital micromirror device (DMD) to obtain Raman spectra ever more efficiently. The advantage in the spatial multiplicity is given by the number of detection channels in a 2D sensor, which can easily achieve a signal enhancement of 10

, however, the enhancement is practically limited by the intensity of the laser that can be incident to the sample.

	One of the advantages of Raman imaging is that it can provide rich information from a sample as a spectrum. In the symposium, there were also discussions about how to efficiently measure thousands of spectra to produce hyperspectral images of biological samples. Especially in coherent anti-Stokes Raman spectroscopy (CARS) and stimulated Raman scattering (SRS) microscopy, techniques for tuning pump or probe wavelength have been developed for the rapid acquisition of SRS or CARS spectra. Yasuyuki Ozeki of the University of Tokyo presented his achievement of rapid wavelength switching using fiber laser and amplification technologies. Spectrum focusing is another technique recently utilized to obtain hyperspectral images in coherent Raman microscopy, where two chirped pulsed lasers realize selective excitation of the vibrational state with a spectral resolution higher than the bandwidth of the laser. Paola Borri of Cardiff University and Ji-Xin Cheng have fully utilized this technique to perform various biomedical applications with CARS and SRS.

	Marcus Cicerone of Georgia Institute of Technology presented his achievement in broadband Raman imaging and provided the comparison of different hyperspectral imaging techniques. The advancement of this coherent Raman technique is associated with the development of laser and detector technologies. Hideaki Kano of Tsukuba University reported broadband CARS microscopy with improved spectral resolution using a nanosecond laser source and supercontinuum, to produce multimodal label-free imaging of cells combined with second harmonic generation (SHG) imaging and third harmonic generation (THG) microscopy.

	With spontaneous Raman spectroscopy, the spectrum is generated by irradiating a single-frequency laser, however, the low Raman scattering cross-section has always hindered rapid image acquisition. This issue has been overcome by the parallel detection of spectra from hundreds to a thousand points in a sample. Fujita pointed out that the efficiency of parallel detection is simply proportional to the number of pixels on a detector, which reaches more than 10

 when using current 2D sensor arrays. Cicerone also mentioned that the difference between coherent and spontaneous Raman microscopy becomes small when a wide spectrum range is measured.

	The use of Raman tags was another topic discussed in many presentations. Although Raman spectroscopy is a label-free technique, its combination with labeling techniques provides innovative approaches to biomedical imaging. In Raman tag imaging, a small molecular structure, such as an alkyne, is used as a tag for target molecules. Raman tags can be used to label small molecules that cannot be labeled by bulky fluorescent probes that may alter the property of target molecules. Mikiko Sodeoka, of the RIKEN research institute in Japan, summarized her achievements in Raman tag imaging, including the first demonstration of alkyne tag imaging and applications to drug, lipid, and organelle imaging. Duncan Graham of the University of Strathclyde and Fujita also discussed the application of alkyne tag imaging in intracellular drug detection with improved sensitivity in SRS microscopy and spontaneous Raman microscopy, respectively.

	The narrow-band emission is another advantage of the Raman tag. The spectrum width of the Raman tag is ten times smaller than that of fluorescence emission. Wei Min of Columbia University utilized this advantage for super multiplexed imaging of intracellular structures with SRS microscopy. By combining pre-resonant excitation, he successfully improved the signal strength and achieved a greater detection sensitivity. The use of the resonance effect usually requires a relatively large tag, but the capability of multiplexed imaging is useful in many areas of biological and medical research when combined with the power of labeling techniques developed for fluorescence microscopy.

	The carbon-deuterium (CD) stretching vibration can also be detected separated from the endogenous molecules by Raman scattering. Lingyan Shi of the University of California, San Diego, reported the use of CD stretching vibration for observation of metabolism in animals. Supplying heavy water to animals, newly synthesized protein, lipid, nucleic acid containing CD vibrations can be observed by SRS microscopy. Narrow emission of luminescence can be combined with Raman imaging. Tao Wu of the Czech Academy of Sciences introduced the use of lanthanide, which provides a narrow band emission at the silent region, for counterstaining cellular structures. The lanthanide probe helps to interpret the Raman spectrum in a sample by highlighting the sample structures.

	The capability of label-free imaging is one of the advantages of Raman imaging methods. During the symposium, there were several reports on implementing Raman microscopy techniques in biomedical diagnosis. Jürgen Popp of the Leibniz Institute of Photonic Technology introduced his group’s long-time efforts to develop a portable coherent Raman imaging system that can be carried into an operating room. The system was realized by a compact fiber laser system that solved a practical bottleneck of the technique. Ji-Xin Cheng developed a handheld coherent Raman imaging system that can be easily operated by a medical doctor during an operation. Mariko Egawa of the Shiseido Global Innovation Center provided several examples of coherent Raman imaging in skin diagnosis and discussed the advantage for studies of cosmetics.

	The combination of Raman imaging with endoscopy is also a promising approach to implement Raman spectroscopy as a diagnostic technique in the clinic, especially combined with robotic surgery. Mamoru Hashimoto of Hokkaido University introduced his group’s recent development of CARS endoscopy using a gradient index (GRIN) lens for image guiding, which can be implemented with such a surgical system. Hervé Rigneault of Institut Fresnel, CNRS, described a technique using a hollow-core fiber that can scan a focused laser in a small region in a body. Hirohiko Niioka of Osaka University introduced the use of a computational technique to obtain clear images with short exposure time, which provides feasibility for the technique for practical applications.

	Cytometry is another direction for utilizing the analytical power of Raman spectroscopy. By combining cytometry with fast imaging using coherent Raman microscopy, large numbers of cells can be analyzed in a short time. Ji-Xin Cheng and Yasuyuki Ozeki presented their recent achievements in coherent Raman cytometry. When utilized with spectral analysis techniques, Raman cytometry can provide many possibilities in cell analysis, such as cell evaluation in medical diagnosis, regenerative medicine, and pharmaceutical studies. Ozeki presented flow cytometry based on fast SRS imaging. Nicholas Smith of Osaka University presented the combination of quantitative phase imaging and spontaneous Raman imaging, which enables rapid diagnosis of cells by using morphological and molecular information provided by optical phase and Raman imaging, providing both the benefits from imaging speed and the powerful analytical power of spontaneous Raman spectroscopy.

	As mentioned above, the spectrum analysis technique is a key element for applying Raman imaging for many biomedical applications. The spectrum contains a large amount of information described as data points in multi-dimensional space, and extracting the information is crucial for practical use. Typically, multivariate analysis, such as principal component analysis and clustering analysis, is utilized, and the adaptation of the techniques for each application has to be performed. Alison Hobro of Osaka University presented a result of extracting spectral components that reflect intracellular components changing with immune response in living macrophage cells. Tamiki Komatsuzaki of Hokkaido University introduced an approach that can merge information analysis and optical measurement to improve the speed of cell and tissue diagnosis dramatically, where the results of spectral analysis are fed back to the measurement condition in real time, allowing an intelligent spectrum measurement for rapid diagnosis. Thomas Bocklitz of the Leibniz Institute of Photonics Technology emphasized the importance of gathering high-quality data together with measurement conditions and information about instruments. The format for such metadata has not been defined, but it is in high demand for practical use of Raman spectra in most applications. The use of machine learning can be one way to mitigate the technical and practical problems, however, this work has just started and we need more experience in technology development and application.

	Raman microscopy is a new modality for biomedical imaging and benefits from the new developments for other optical imaging techniques. Wei Min demonstrated 3D SRS imaging of deep tissue using a tissue clearing technique, enabling a 10-fold increase of depth penetration. Zhiwei Huang of the National University of Singapore described his new approaches for achieving super-resolution in coherent Raman imaging. Several new efforts to improve the speed and signal, such as a technique using Fourier-transform spectroscopy, were also discussed.

	During the symposium, a panel discussion entitled “The Future of Raman Imaging” was also held. Rachel Won of the journal 

 also joined the panelists and provided perspectives from broader aspects of the related science and technologies. We discussed current issues of Raman imaging techniques and requirements for shaping the many areas of potential of the technique into “real” biomedical applications.

	Defining a data format, as mentioned above, and standard procedures to certify the reliability of the measurement, are also needed. To do this, it would be necessary to have a core facility where different Raman imaging techniques, equipment for biomedical studies, and data analysis are available for interdisciplinary collaboration.

	The two-and-half day symposium gave us a strong impression of the advance of Raman imaging technology and the expectation for biomedical applications that provide new approaches for making the invisible visible. The symposium was co-chaired by Eiichi Tamiya and Katsumasa Fujita of Osaka University and the Advanced Institute of Science and Technology (AIST) and Prabhat Verma of Osaka University.

	This wonderful symposium was made possible by support from AIST-Osaka University Advanced Photonics and Biosensing Open Innovation Laboratory (PhotoBIO-OIL), JSPS Core-to-Core Program, Photonics Center in Osaka University, and The Spectroscopical Society of Japan, along with the sponsors. The chairs also thank the international steering committee, Duncan Graham, Wei Min, Jürgen Popp, and Nicholas Smith, for their contribution to gather the excellent speakers and the enthusiastic participants. We are preparing a special issue in 

, a journal from the Royal Society of Chemistry, and have started planning the next symposium.

is with the Department of Applied Physics at Osaka University, in Osaka, Japan. Direct correspondence to: 

Recent technical advances in biomedical Raman imaging pave a way to its application in the biomedical fields, where morphological information of samples provides rich information. A recent technical conference in Osaka, Japan, explored these developments.

Katsumasa Fujita

Articles

A Further Leap of Biomedical Raman Imaging

Biomedical Raman Imaging 2019 in Osaka