A crucial step to treating disease is first understanding how the disease works. To do this, scientists need to get a good look at the roots of the disease—something that is easier said than done. Imaging deeply into a living organism—be it a cell, a fish, a human, etc.—has always been a bottleneck for biologists, even with the invention of fluorescent tags, Raman microscopy and a host of non-destructive imaging techniques.
Fluorescence hyperspectral imaging, or fHSI, is the most commonly used technique among biologists to peer inside an organism. The method can differentiate colors across a spectrum, tag molecules so they can be followed and produce vividly colored images of an organism's insides. However, data acquisition is a problem. Since fHSI produces so much data, due to the complexity of biological systems, the data is analyzed after acquisition, leaving an interval in the process timeline without information. In certain scenarios, the “late” data can kill research or medical results.
To fill this gap, researchers at the University of Southern California developed a novel algorithm called spectrally encoded enhanced representations (SEER) that is focused on enhancing the visualization of hyperspectral data. The algorithm performs up to 67 times faster and at 2.7-fold higher accuracy than present techniques.
“SEER exploits our previously developed denoising algorithms and combines them with a novel set of adaptive ‘Fourier space’ color maps to enhance the visualization of small color differences,” lead author and USC professor Francesco Cutrale explained to Laboratory Equipment. “This pre-processing step is meant as a rapid interface for any user with the content of information of hyperspectral technology.”
The key to SEER lies in its ability to detect subtle changes, which are often important tells within biological systems. Current methods focus on interpreting data the same way our eyes would perceive it, while SEER can process vibrant fluorescent tags across the full spectrum of colors, discriminating very subtle color and spectral differences between samples.
For example, the first application of SEER is detecting early stages of lung disease and potential damage from pollutants.
“[Our study] showed rapid visualization of differences in metabolism in cells of the airway, corresponding to different roles of cells. The metabolic readout can be informative of the overall health of the tissue, for example, when affected by external factors such as pollution or disease,” said Cutrale.
In addition to this lung disease research in collaboration with doctors at Children's Hospital Los Angeles, scientists in the life sciences field have started adopting SEER in their experimental pipelines in an effort to further improve efficiency.
Cutrale said he and his team are also interested in expanding toward digital and virtual medicine. The researchers believe SEER is versatile enough to be integrated with mobile phones in the next few years, possibly in the form of an app.
“SEER would provide a powerful visualization tool for applications like remote medicine, food safety or counterfeit detection,” said Cutrale.
The algorithm was first authored by Wen Shi and Daniel Koo at the Translational Imaging Center of USC.
Photo: New spectrally encoded enhanced representations (SEER) technology produces sharper images of zebrafish and other organisms quicker than existing methods. Credit: Wen Shi, Daniel E.S. Koo and Francesco Cutrale of USC