Stimulated Raman scattering (SRS) microscopy is a label-free quantitative chemical imaging technique that has demonstrated great utility in biomedical imaging applications ranging from real-time stain-free histopathology to live animal imaging. However, similar to many other nonlinear optical imaging techniques, SRS images often suffer from low signal to noise ratio (SNR) due to absorption and scattering of light in tissue as well as the limitation in applicable power to minimize photodamage. We present the use of a deep learning algorithm to significantly improve the SNR of SRS images. Our algorithm is based on a U-Net convolutional neural network (CNN) and significantly outperforms existing denoising algorithms. More importantly, we demonstrate that the trained denoising algorithm is applicable to images acquired at different zoom, imaging power, imaging depth, and imaging geometries that are not included in the training. Our results identify deep learning as a powerful denoising tool for biomedical imaging at large, with potential towards in vivo applications, where imaging parameters are often variable and ground-truth images are not available to create a fully supervised learning training set.

Intraoperative consultations, used to guide tumor resection, can present histopathological findings that are challenging to interpret due to artefacts from tissue cryosectioning and conventional staining. Stimulated Raman histology (SRH), a label-free imaging technique for unprocessed biospecimens, has demonstrated promise in a limited subset of tumors. Here, we target unexplored skull base tumors using a fast simultaneous two-channel stimulated Raman scattering (SRS) imaging technique and a new pseudo-hematoxylin and eosin (H&E) recoloring methodology. To quantitatively evaluate the efficacy of our approach, we use modularized assessment of diagnostic accuracy beyond cancer/non-cancer determination and neuropathologist confidence for SRH images contrasted to H&E-stained frozen and formalin-fixed paraffin-embedded (FFPE) tissue sections. Our results reveal that SRH is effective for establishing a diagnosis using fresh tissue in most cases with 87% accuracy relative to H&E-stained FFPE sections. Further analysis of discrepant case interpretation suggests that pseudo-H&E recoloring underutilizes the rich chemical information offered by SRS imaging, and an improved diagnosis can be achieved if full SRS information is used. In summary, our findings show that pseudo-H&E recolored SRS images in combination with lipid and protein chemical information can maximize the use of SRS during intraoperative pathologic consultation with implications for tissue preservation and augmented diagnostic utility.

Optical “virtual biopsy” is an attractive way to improve disease diagnosis and surgical guidance. Many optical microscopy techniques have been developed to provide diagnostic information without the need for tissue sectioning or staining. Among these techniques, label-free chemical imaging is the most desirable. Recently, it has been shown that narrowband, picosecond stimulated Raman scattering (SRS) can achieve comparable morphological contrast to hematoxylin and eosin staining (H&E staining), the ‘gold standard’ of pathology. However, to translate the technique from the bench to the bedside, optimal laser sources and parameters have yet to be identified. Here we describe an improvement to the narrowband SRS microscopy techniques for label-free tissue imaging. Through spectral slicing of broadband, femtosecond pulses, we are able to maintain the same protein/lipid contrast as narrowband SRS while achieving a higher signal-to-noise ratio (SNR). Our method draws upon the benefits of femtosecond pulses (e.g. higher peak power) while preserving those of picosecond pulses (e.g. adequate spectral resolution). We demonstrate this achievement through protein/lipid signal and contrast quantification of mouse brain tissue as a function of bandwidth, and comparison with numerical simulations. Further method validation is provided through imaging of additional mouse tissues: liver, kidney, and skin.

Fingerprints have long been the gold standard for personal identification in forensic investigations. Methods for cultivating and enhancing the visualization of latent fingerprints (LFPs) are continuously evolving. One important challenge is to identify suspicious chemicals present in fingerprint residues, which requires chemical imaging capability. Recently, vibrational spectroscopy has shown that LFP analysis through tape-lift, Raman mapping, and multivariate data analysis presents a useful tool for forensic investigation. However, there are still major difficulties in terms of acquisition speed, poor spatial resolution, and lack of sensitivity. This paper demonstrates the feasibility of stimulated Raman scattering microscopy to quickly and easily extract LFP patterns from different substrates. Contrary to what has been reported, no obvious fingerprint degradation or lipid diffusion is observed with either glass or stainless steel substrate. Importantly, we demonstrate that trace exogenous chemicals can be detected in fingerprints. We further demonstrate an improvement in directly acquiring a LFP pattern lifted from tape by spectrally removing signals from tape.