Researchers from the University of Illinois at Urbana-Champaign have added infrared capability to a standard optical microscope in the hopes of digitalizing cancer diagnoses. The scientists were able to use infrared measurements in conjunction with machine learning algorithms to create digital biopsies that were very similar to biopsies done with traditional methods.
Lead researcher and professor of bioengineering and the director of the Cancer Center at Illinois, Rohit Bhargava, commented, “The advantage is that no stains are required, and both the organization of cells and their chemistry can be measured. Measuring the chemistry of tumor cells and their microenvironment can lead to better cancer diagnoses and better understanding of the disease.”
The traditional methodology is to add a stain to the tissue so that pathologists can see the shapes of the cells when studied under a microscope. It is sometimes hard to distinguish cancer cells from healthy ones or to determine the boundaries of a tumor, meaning that the diagnosis is subjective and may not tell the whole story.
"For more than a century, we have relied on adding dyes to human tissue biopsies to diagnose tumors. However, the shape and color induced by the dye provide very limited information about the underlying molecular changes that drive cancer," Bhargava said.
Bhargava, and associates, created a hybrid microscope by using a specialized lens and an infrared laser with an optical camera. This hybrid microscope can measure infrared data and a high-resolution image. First paper author and postdoctoral fellow, Martin Schnell, said, “We built the hybrid microscope from off-the-shelf components. This is important because it allows others to easily build their own microscope or upgrade an existing microscope.”
The researchers verified their results by imaging both healthy and cancerous breast tissue samples. They compared the hybrid microscope’s digital “stain” with ones from the traditional technique. The digital biopsy directly correlated to the traditional one.
"Infrared-optical hybrid microscopy is widely compatible with conventional microscopy in biomedical applications," Schnell said. "We combine the ease of use and universal availability of optical microscopy with the wide palette of infrared molecular contrast and machine learning. And by doing so, we hope to change how we routinely handle, image and understand microscopic tissue structure."
The researchers published their results in the Proceedings of the National Academy of Sciences.