AccessPath is an affordable system for immediate digital pathology of resected tumors. Tissue specimens are stained with inexpensive vital dyes and illuminated with UV excitation, localizing fluorescence emission to a thin surface layer. A custom optical phase mask and associated deep-learning algorithm jointly operate to extend depth-of-field (DOF) by >10X, resolving subcellular features in a DOF of 200 µm, allowing intact tissue specimens to be rapidly scanned without serial refocusing.

qOBM is a recently developed label-free optical imaging technology that enables quantitative phase imaging and 3D refractive index tomography of thick tissues. The technology achieves high contrast imaging of cellular and subcellular structures to enable real-time ex and in vivo label-free and slide-free histology. In this presentation, I will describe the working principles of qOBM and show its potential for POC histology.

In order to catalyze a digital pathology transformation to improve clinical decisions and patient outcomes, we are developing a novel technological approach that offers significant advantages over traditional “gold-standard” histopathology in terms of accuracy and throughput. First, we have developed an open-top light-sheet (OTLS) microscopy platform for slide-free 3D pathology of large clinical specimens, enabling whole biopsies and surgical specimens to be non-destructively imaged with an ease of use similar to a flat-bed document scanner. Next, using both hand-crafted and deep-learning AI techniques, we are quantifying 3D spatial and molecular biomarkers for enhanced clinical decision support, such as for prognosticating patient outcomes (indolent vs. aggressive disease) and to guide resection procedures.

Structured illumination microscopy (SIM) has a number of compelling applications in direct-to-digital onsite pathology, including large-area rapid 2D imaging for core biopsy and whole-resection tumor margin assessment, and rapid nondestructive cytologic imaging. Additionally, the combination of emerging concepts in computational microscopy, image analysis, and artificial intelligence may accelerate or ease translational burdens for these technologies. We will discuss these topics and more in the context of point-of-care histology for diagnosis and treatment of prostate cancer.

Our lab at UC Davis has developed a new imaging method, FIBI (Fluorescence Imitating Brightfield Imaging), for the histopathological assessment of freshly excised or fixed specimens. This method generates diagnostic-grade images without requiring previous preparation of a glass slide. FIBI images are generated by virtual back-illumination of a stained tissue surface of stained tissue, using 405 nm LED illumination configured in epifluorescence mode and conventional microscope objectives. The resulting images immediately resemble familiar glass-slide-based brightfield histology. We will discuss this technology's utility for various applications, including rapid onsite evaluation (ROSE), point-of-care diagnostics, intraoperative margin assessment and global deployment.

Reflectance confocal microscopy images cellular patterns and morphology in human skin and detects skin cancers with high sensitivity and specificity. Following two decades of development and clinical trials, RCM imaging of skin was granted reimbursement codes in 2016. The imaging is now being adopted and advancing into routine use for noninvasively guiding diagnosis, treatment, and management of skin cancers. This talk will present a critical overview of this field.

Methods for intra-operative determination of surgical margins via ex-vivo thick tissue analysis could lead to more successful cancer surgeries. Multiple labeled and label-free microscopy methods have been proposed with this objective, yet none has emerged with sufficient histological realism, speed, and utility to change clinical practice. I will discuss recent developments in bringing ultraviolet absorption, scattering, and autofluorescence emission contrast into a single imaging platform. Absorption imaging is achieved using photoacoustic imaging, photoacoustic remote sensing, or negative contrast in autofluorescence imaging. Autofluorescence in multiple optical bands enable simultaneous imaging of tissue optical redox ratio for mapping metabolic activity, co-registered with virtual histology. These developments, coupled with advances in deep learning, have enabled realistic virtual histology capabilities with clinically acceptable sensitivity and specificity compared to gold-standard H&E histology. Moreover, scan rates as high as 7min/cm2 with 400nm resolution were achieved with room for significant speedups in future work.

From expedited diagnoses to enhanced patient care, AI presents revolutionary prospects in point-of-care histology. However, challenges such as scarce labeled data and variations across devices and facilities must be addressed for robust solutions. In this talk, I’ll delve into some promising uses for AI, demonstrate the primary obstacles that medical device developers should be prepared for, and propose solutions that can enable AI to transform point-of-care histology practices.