Stimulated Raman Histology (SRH) allows imaging of samples based on intrinsic molecular contrast without the need for sectioning and staining required by traditionally histologic methods. As such it is highly suitable for rapid on-site assessment of fresh tissue biopsies at the time of the procedure. Combination with artificial intelligence (AI) based image interpretation has the promise to provide an end-to-end solution for near real-time histopathology. SRH has been utilized in 5000+ cases across brain, lung, prostate, breast and pancreatic cancer. Here we review the development, clinical translation and future directions of SRH.

Despite the approval of immunotherapy in 140 indications, and some patients showing great response, only 24-40% of treated patients respond, suggesting single markers such as PD-L1 are insufficient to match patients to therapy. There’s a “call to action” to shift away from classifying tumors by organ and instead focus on biomarker expression patterns. Only with scalable technology platforms that comprehensively probe biomarker status, diagnostic subclassification, immune status, and the tumor microenvironment, can we hope to understand the underlying disease biology and consequently effectively match patients to the best treatment regime. Further, other regions such as the extracellular matrix (ECM) may yield complementary information, as collagen-rich ECMs may present a barrier to drug diffusion and the orientation of fibers may direct the migration of malignant cells. Here we present a deep-learning based virtual staining method, using whole-slide autofluorescence imaging, which transforms a single unstained tissue section into a panel of perfectly registered virtual biomarker and morphology assays. Further we demonstrate computational multiplexing of results and the ability to combine with chemical methods such as immunohistochemistry or spatial transcriptomics. In this talk we will explore applications of the technique to identify and understand spatial expression patterns, which may represent novel phenotypic signatures of disease.

Deep learning techniques create new opportunities to revolutionize tissue staining methods by digitally generating histological stains using trained neural networks, providing rapid, cost-effective, accurate and environmentally friendly alternatives to standard chemical staining methods. These deep learning-based virtual staining techniques can successfully generate different types of histological stains, including immunohistochemical stains, from label-free microscopic images of unstained samples by using, e.g., autofluorescence microscopy, quantitative phase imaging (QPI) and reflectance confocal microscopy. Similar approaches were also demonstrated for transforming images of an already stained tissue sample into another type of stain, performing virtual stain-to-stain transformations. In this presentation, I will provide an overview of our recent work on the use of deep neural networks for label-free tissue staining, also covering their biomedical applications.

We present the product development of a Photon Absorption Remote Sensing (PARS) based Advanced Imaging Device (AID) for label-free histological imaging in the field of Life Sciences and Drug Development. System level form factor, system architecture, system operation, and product features, including multiple histochemical virtual stains of human tissue, are described. An alpha prototype of the AID will be online in 2024 followed by beta prototypes and the subsequent AID product launch.

A Photon Absorption Remote Sensing (PARS) microscope is presented for label-free histological imaging with high-resolution, and high-sensitivity, with pragmatic imaging speeds 10x faster than previous embodiments compatible with intraoperative diagnostics. PARS directly measures endogenous scattering and optical absorption (radiative and non-radiative) at multiple wavelengths, capturing all dominant light matter interactions simultaneously. When applied to unstained human tissue specimens, PARS may directly reproduce histochemical staining (e.g., H&E staining) through AI based virtual staining algorithms. Concurrently, PARS unique contrasts may directly facilitate novel chromophore specific visualizations.

A third of all food produced for human consumption is wasted from the farm to the fork. FAO’s Sustainable Development Goal Target 12.3 aims at reducing by half per capita food waste by 2030. Using hyperspectral measurements, chemometric approaches and computational algorithms including AI, we’ve developed a simple but robust methods allowing prediction of the remaining shelf life of fresh produce. This information will allow actors along the food chain to optimize produce usage, and reduce food waste.