The staining of cells has been a part of the pathology workflow for decades, allowing clinicians and researchers to identify different cell types and to track their activity. As any laboratorian knows, it can be a slow and laborious process and can also cause disruption to the sample being analyzed. At the University of Illinois at Urbana-Champaign, researchers have developed a virtual staining technique that doesn’t require any actual staining while producing similar results.
They’re using infrared spectroscopic imaging to analyze the chemical composition of cells directly. By noting the spectral characteristics of light bouncing off the cells, the investigators were able to identify correlations that a computer can use to do quick pathology analysis. The researchers hope that the new method will drastically increase the speed of digital pathology studies, since manual aspects of staining and preparation may be avoided altogether.
According to Rohit Bhargava, PhD, professor of chemistry and engineering at University of Illinois at Urbana-Champaign and author of the study, the move to stainless staining makes perfect sense from an efficiency standpoint.
“We’re in the midst of a trend where we’re becoming more molecularly sensitive,” he noted. “We’re using more information from limited amounts of tissue. Many times, the tissue, animal model or needle biopsy is limited. That’s not to mention limitations surrounding time and cost, as human and materials costs increase with number of stains. Finally, it may not be possible to repeat staining and wash again because of limited material or changing tissue structure.”
Sectioning Once, Staining Forever
Although the timing was perfect to obtain cellular information in chemistry, Bhargava said another strong advantage of stainless staining is the ability to translate the image into a format that’s easy to understand for practitioners.
Instead of using stains, the spectra measures the chemical constitution of cells and tissues directly. By using computational techniques, the researchers were able to relate spectral properties to known staining patterns of tissue. Therefore, molecular stains can be reproduced without staining the tissue, but by using the intrinsic molecular contrast of the tissue and computation. So, in addition to not accruing unnecessary material cost, time or effort, this method also leave precious tissue pristine for downstream analyses, according to the research team.
Another application of the approach can be in the analysis from a thin needle biopsy, or other small amounts of samples. In cases where materials are limited or there may be a need to closely correlate multiple expressed molecules; it may not be possible to obtain multiple samples from the same biopsy for multiple stains. According to Bhargava, the method developed in this study could be a solution, allowing the user to simply “dial-in” a required stain. The study is timely as it builds on the emergence of chemical imaging and maturation of computation from the sciences/engineering side and the drive to greater molecular content from the biomedical/clinical side.
The study utilizes computation, instead of staining to provide images.
SEE ALSO: Histopathology
“We’re relying on the chemistry to generate the ground truth and act as the ‘supervisor’ for a supervised learning algorithm. A human has to verify that the stain was applied appropriately, but that’s something that pathologists do very well,” said David Mayerich, PhD, lead author of the study, assistant professor at the University of Houston. “One of the bottlenecks in automated pathology is the extensive processing that must be applied to stained images to correct for staining artifacts and inconsistencies. The ability to apply stains uniformly across multiple samples could make these initial image processing steps significantly easier and more robust.”
This study illustrates the closer connections between imaging technology and clinical applications. “Infrared and optical imaging seemed to be distinct modes for getting important data in pathology. This study shows a close link between the two, allowing the user to choose the best method to address their needs,” commented Bhargava.
Admittedly, stainless staining requires an initial investment in desktop software that reveals the exact tissue dyes; Bhargava said one of their goals was ease of use so the instrumentation is similar to that of any microscope and the software features regular drop-down menus.
