Raman spectroscopy and imaging are powerful, label-free techniques for a wide range of medical applications. By enabling the identification of molecular composition and revealing the structural information of biological tissues and fluids with high specificity, Raman methods help clinicians and researchers to identify pathological changes — often before they become visible via traditional imaging. The evolution of Raman spectrometers and microscopes from bulky, laboratory-bound instruments to compact, portable, and clinically viable tools at the point of care represents a major technological shift in optical spectroscopy. Today, portable Raman spectrometers and microscopes offer high-resolution, real-time molecular analysis in both field and clinical settings. Breakthroughs in laser miniaturization, advancements in optical design, the development of compact, high-sensitivity detectors, and the use of dedicated optical probes have combined to enable the creation of smaller, more robust systems without compromising performance. In addition to diagnostics, the availability of these Raman instruments is enabling intraoperative tissue characterization and image-guided interventions. Moreover, their integration with wireless communication, cloud connectivity, and AI-powered data interpretation further enhances usability and remote diagnostics and decentralized workflows. Yet despite the vast potential of Raman spectroscopy and imaging in the medical sector, and even as progress continues in the technology space, the adoption of Raman methods in clinical environments faces several significant challenges. Independent of some technological limitations, translating Raman instruments from laboratory to clinical use demands rigorous standardization, validation, and regulatory approval, which can be time-consuming and costly. Also, the complexity of spectral data often necessitates the use of advanced algorithms and trained personnel for accurate analysis. Access to funds can also influence the evolution of the development and its ultimate application. From within industry, companies are rising to the challenge of harnessing Raman’s capabilities. The innovative approaches these companies are deploying showcase a convergence of sophisticated instrumentation and high-performance solutions. Tumor margin identification As the demand grows for personalized and precise medical care, Raman-based technologies are playing an increasingly important role in diagnostics and therapeutic monitoring. This role is poised to expand even further as advanced instrumentation enters clinical settings and the marketplace. Leveraging the combination of Raman spectroscopy and microscopy, the company Refined Laser Systems, based in Münster, Germany, has developed an alternative approach to identify tumor margins directly in the surgery room. The company’s solution is based on an intraoperative microscope. It aims to provide pathologists and surgeons with an image of tissue — within minutes — that has sufficient contrast for clinical decision-making during surgical procedures. Refined Laser Systems’ device is based specifically on a nonlinear variant of Raman scattering called stimulated Raman scattering (SRS). This third-order nonlinear phenomenon involves a second photon — called the Stokes photon — which stimulates a specific transition. When the difference in frequency between both photons resembles that of a specific vibrational or rotational transition, the occurrence of this transition is resonantly enhanced. Refined Laser Systems’ technology streamlines what is considered today to be the gold standard approach for cancer diagnostics. It addresses the repeated staining of biopsied tissue (initially with hematoxylin and eosin, followed by immunohistochemical staining), which requires time-consuming preparation steps. Tissue removal, thin sectioning, and staining must all take place before a pathologist can make a diagnosis. Despite many advancements, this process typically takes up to 10 h, prolonging medical procedures. Additionally, a recent report from the German Cancer Society stated that an average of 12% of breast cancer surgeries in Germany must be repeated because not all cancer cells are detected and removed during the first surgery. In real-world use, the SRS technique can accelerate image acquisition by a factor of a thousand, potentially enabling its use for intraoperative assessments. Plus, it has been previously shown that SRS can provide label-free contrast to differentiate cell nuclei and cell bodies, both of which are key features of currently applied hematoxylin and eosin-based histopathology (Figure 1). By quantifying the ratio of CH2 and CH3 chemical bonds found in cell bodies and nuclei, respectively, microscopic images of healthy and cancerous tissue sections can be generated in near-perfect concordance with standard histology. Figure 1. A label-free virtual hematoxylin and eosin-stained liver section. Courtesy of Refined Laser Systems. Refined Laser Systems has advanced its fiber-based laser technology to develop a system for the application of SRS in the field, offering rapid and wide tuning capability for the analysis of complex samples. This advancement has supported a move from the complex laboratory setups that the company used in previous studies toward a more compact and integrated solution. With Refined Laser Systems’ laser technology, it is now possible not only to generate histological images but also to efficiently access the entire Raman spectrum as a source of novel biomarkers. This means that a user can establish new biomarkers that could, for example, predict the outcome of immunotherapies. This innovative Raman-based diagnostic platform brings the power of multidimensional tissue analysis directly to the patient’s bedside. By using the full vibrational spectrum of tissue samples, it delivers rich histological data without the need for invasive procedures or extensive lab processing. Lab-to-clinic challenges Following a similar approach, the company Cambridge Raman Imaging has successfully developed the CORAL (COherent RAman pLatform) system. This multimodal, label-free imaging platform offers subcellular resolution and enables the capture of high-resolution images leveraging broadband SRS in the CH-stretching region (2800 to 3100 cm−1). It integrates Cambridge Raman Imaging’s laser architecture, which passively synchronizes two fiber oscillators, with an ultralow-noise multichannel detector. The mechanism is specifically developed for the application of Raman medical imaging and detection in a multiplex SRS microscopy setup. Further, Cambridge Raman Imaging’s advanced laser system delivers both pico- and femtosecond dual-output pulse durations. This facilitates access to a broad range of nonlinear optical microscopy techniques, including two-photon excitation fluorescence (TPEF), second-harmonic generation (SHG), and pump-probe. The multichannel detector of the CORAL system is a lock-in amplifier spectrometer that the company engineered for single-shot CH-stretching detection, offering 38 spectral channels, each with an independent lock-in amplifier. This capability enables the system to disentangle the paraffin contribution from that of the lipids, which overlap in this region, so that imaging can be performed for both fresh or newly frozen tissue samples as well as for formalin-fixed paraffin-embedded samples. The system’s compact design allows real-time average power and demodulated signal extraction in parallel on all channels. The system also collects two more signals — TPEF and SHG — simultaneously through epi-detection. This supplements the 38 channels collected by the multichannel lock-in amplifier and provides comprehensive information about the organic samples under analysis. With this integrated solution, Cambridge Raman Imaging delivers detailed images across various cytological and histological sample types. As mentioned, the transition of this kind of system from the lab to a clinical environment demands careful attention to hardware, software, safety, and design. The necessary focus that a company must devote to each of these areas can strain a manufacturer tasked with making the changes needed to adapt the solution to its final use without compromising its effectiveness. In the case of the CORAL system, Cambridge Raman Imaging has targeted its hardware: The company drastically reduced and ruggedized all optical and mechanical components so that the system fits comfortably in chemistry labs and even operating theaters. Tool-free access panels and modular mounts allow rapid alignment and part replacement. And built-in self-calibration routines and remote diagnostics minimize downtime and eliminate the need for onsite specialists. The company has also prioritized the assurance of safety, enclosing the system’s Class 4 laser in a fully interlocked Class 1 housing. Meanwhile, impact- resistant windows, redundant interlock circuits, and automatic shutters eliminate the risk of accidental exposure. Finally, Cambridge Raman Imaging designed and refined its system software with clinical users to replace complex, lab-style controls with a simpler, user-friendly workflow. It also added the possibility for users to inspect the Raman spectrum at each pixel and watch image construction. The CORAL system can also generate “virtual stains,” which closely resemble conventional hematoxylin and eosin slides, complete with color-coded overlays marking healthy and tumor areas. Improving light throughput The inherently weak Raman spectroscopy signal — which can require longer acquisition times or the use of higher laser powers — is a technical limitation of Raman-based techniques. This obstacle is especially detrimental to certain applications, including those in biomedicine. To resolve this limitation, Thorlabs, based in New Jersey and with European offices in Germany, France, England, and Sweden, designed Raman spectrometers that replace the conventional single-slit input apertures with a defined pattern of multiple slits. This design concept uses a pseudo-Hadamard mask of order 64 as the aperture (called a coded-aperture, or CODA input), and it acts as a convolution operator with an orthogonal basis in two dimensions1. Hadamard matrices consist of elements that take only the values of 1 and −1. Since the −1 element cannot be realized in a real-world mask, Thorlabs’ pseudo-Hadamard optical mask maps the −1 to 1 values to intensity modulations ranging from 0 (light blocked) to 1 (pixel transparent). The resulting mesh image from the sensor is reconstructed using the inverse transform operator (Figure 2). Figure 2. A size comparison of the coded aperture and a standard slit aperture, an innovation from Thorlabs. The coded aperture is a pseudo-Hadamard mask of order 64. The white and dark areas represent regions where light is transmitted through the mask or blocked, respectively. Courtesy of Thorlabs. Through the combination of the pseudo-Hadamard mask with an aperture size of 2.3 mm × 3.2 mm, Thorlabs improved light throughput enough for sufficient Raman-scattered light to reach the CMOS sensor, enabling room temperature operation. The increased étendue of the spectrograph’s input also enables a larger spot size (Ø1 to Ø2 mm), which allows sample excitation with an unfocused laser beam. This is ideal for nonhomogeneous samples, such as those found in certain biomedical applications, which require averaging over a large area to determine composition. An unfocused excitation beam also results in a lower power density, which reduces the risk of laser-induced damage to the sample. Thorlabs’ spectrometers have found several interesting biomedical applications. These include the characterization of calcified tissue in biological samples, the measurement of the spectral shifts of certain amino acid vibrations upon the addition of markers, and the distinguishing between different beta-blockers, such as metoprolol salts. Device innovation and financial hurdles Securing funding and navigating complex regulatory pathways remain two of the biggest challenges for companies developing medical devices, often turning groundbreaking innovation into a race against time, resources, and compliance. As it relates to regulation and legislation, starting a (pre)clinical study of a new technology can be as demanding as bringing a medical device to market. At the same time, finding the right partner to support development is critical, due to the need for sustained investment over several years. For some companies facing particular difficulties in securing funding, the risk can be a slowdown in development, limited scalability, or a halt to innovation altogether. RiverD International BV, headquartered in the Netherlands, has considerable experience in developing devices for biomedical applications. Its innovations range from its successful in vivo skin analysis product line to its development portfolio of potential medical diagnostic devices. One of these developments is a dedicated Raman spectroscopic device, called MarginGuide, for assessing tumor resection margins on resected tissue (Figure 3). MarginGuide uses a thin optical fiber needle to measure the distance between a resection surface and the tumor in a few seconds, reporting the locations of incomplete tumor resection. This process enables fast and thorough examination of all resection surfaces while the patient is still in the operating room, providing the surgeon with the option to remove additional tissue to achieve a complete resection. Figure 3. RiverD’s Raman system, which has found application in assessing soft tissue resection margins. Courtesy of RiverD International. The development of MarginGuide began as part of an effort to help improve oral cancer surgery — a field for which there is clear need for solutions. RiverD successfully developed the MarginGuide technology with public funding, but it has been unsuccessful in finding the type of private and venture capital funding needed to propel it further toward clinical implementation. Fundamentally, although there is need for the MarginGuide technology in oral cancer surgery, investors often prefer to target solutions for cancers with higher incidence, regardless of competition from other emerging technologies. To overcome this lack of funding, RiverD launched the European Union-funded Spectra-BREAST project last year. The project will enable RiverD to adapt the MarginGuide technology to intraoperative margin assessment in breast cancer surgery. Simultaneously, in collaboration with professor Senada Koljenovic, head of pathology at the University Hospital Antwerp in Belgium, RiverD has also commenced the investigation of its technology in prostate tumor surgery. The company and hospital have also engaged in another project, which is focused purely on the clinical implementation of MarginGuide technology in oral cancer surgery. Members of the RiverD management team understand that making an investable business case without making unrealistic promises or projections is a challenge. In response, the company has favored a slower but steady path. The result is significant support from various organizations to achieve multiple goals. Reference 1. M. Gehm et al. (2006). Static two-dimensional aperture coding for multi-modal, multiplex spectroscopy. Appl Opt, Vol. 45, No. 13, pp. 2965-2974.