Richard Levenson, MD
UC Davis Health
New technologies for tissue microscopy, new opportunities for computational pathology
Background: Light microscopy of tissue biopsies remains the central technique in diagnosis and management of cancers as well as other diseases. The brightfield (transmission) optical design of today's clinical microscopes, which reached near-contemporary performance as far back as the 18th century, requires optically thin slices of tissue mounted onto glass slides. However standard formalin- and paraffin-based histology required to prepare thin sections is a labor- and time-intensive process, and sectioning can consume small specimens, leaving little or nothing for downstream molecular studies. There are alternative microscopy techniques that can directly image thick, unfixed tissue specimens. These includes structured illumination, conventional reflectance and fluorescence confocal microscopy, multi-photon confocal imaging, spectrally encoded confocal microscopy, optical coherence tomography, and photoacoustic approaches, among others, described below. Separately, we have developed methods for improving image data that can be obtained from conventional H&E-stained slides.
Methods: Our laboratory is developing two novel approaches to histopathology imaging: FIBI, for slide-free, direct-to-digital microscopy, and DUET, for use with existing H&E-stained slides. FIBI (flexibly induced backlight imaging) is a non-destructive technique that captures high-resolution microscopy images directly from the face of tissue specimens. It is based on < 1 min H&E staining followed by epifluorescence-based imaging. DUET (dual-mode emission-transmission) generates both brightfield and long-pass fluorescence data from H&E-stained slides using a scanner that generates inherently pixel-matched images. A variety of parametric and machine-learning methods can then be used to display collagen, elastin and basement membrane distribution, as well as other components not easily distinguished by brightfield-only methods, all without requiring the preparation of additional tissue sections and or staining steps.
Results: FIBI: Whole-slide-like images can be captured within a few minutes and directly resemble conventional brightfield histology. Color normalization (optionally AI-based) can be used to enhance the resemblance. At the same time, images also contain additional features that includes surface topology, greater continuity of linear structures such as blood vessels, and, contrast that can highlight macromolecular species that are not well visualized with H&E alone. DUET can generate images from H&E-stained slides that highlight collagen distribution with results that resemble and indeed improve upon trichrome stains as well as second-harmonic generation techniques. FIBI is non-destructive and can be followed by standard FFPE for long-term storage and/or submission for downstream molecular assays. With both FIBI and DUET time and money can be saved, but in addition, enhanced tissue constituent contrast could drive enhanced performance for computationally based analytics. In the meantime, FIBI can providing near-patient, near-real-time histology results that can be used in ROSE (rapid on-site evaluation) of biopsy quality, frozen-section replacements for intraoperative surgical guidance, and point-of-care tissue diagnostics for remote settings.
Conclusions: Novel imaging methods can make good partners with AI-based analysis methods, simultaneously reducing time and expense and potentially improving AI results due to avoidance of some common artifacts in regular histology as well as inclusion of novel contrast.
Richard Levenson, MD, FCAP, is professor and vice chair for strategic technologies, Department of Pathology and Laboratory Medicine, UC Davis Health. He received his MD at University of Michigan and pathology training at Washington University, followed by a cancer research fellowship at Univ. of Rochester and faculty positions at Duke and Carnegie Mellon. He joined Cambridge Research & Instrumentation, becoming VP of Research before assuming his present position at UC Davis. He helped develop multispectral microscopy and small-animal imaging systems, birefringence microscopy, multiplexed ion-beam imaging (MIBI), and slide-free as well as enhanced-content microscopy approaches, and is an inventor on some 15 patents. He is section editor for Archives of Pathology and is on the editorial board of Lab. Invest. and AJP. Regrettably, he also taught pigeons histopathology and radiology. He is a recipient of the 2018 UC Davis Chancellor’s Innovator of the Year award and is a Fellow of SPIE.