Scientists at Harvard University have developed a noninvasive imaging technique that captures images at the molecular level so quickly that they can “watch” red blood cells move through the capillaries of a live mouse. The system uses two laser beams set at different frequencies to excite specific types of molecules in the skin. A custom-designed detector picks up the excited molecular signal and translates it into an image.
Sunney Xie, professor of chemistry and chemical biology at Harvard, says the technique could be a noninvasive alternative to often painful and time-consuming skin biopsies.
“To identify a solid tumor, tumor margins, and metastasis requires cutting and slicing tissue, staining it with dye, and looking at it under a microscope in a pathology lab next door—a process that could take 15 to 20 minutes,” says Xie. “Here, we don’t need a biopsy; we can obtain almost identical images without cutting the tissue.”
Currently, systems like magnetic resonance imaging (MRI) and positron emission tomography (PET) serve as windows into the molecular world. Clinicians use these tools to identify diseases like cancer. To detect specific molecules or cancerous cells, MRI requires the patient to ingest or inject contrast agents, and PET requires low doses of radioactive substances. However, scientists have found that these compounds, also referred to as “labels,” may harm or alter normal cellular processes.
In contrast, Xie’s technique is label-free, drawing upon a noninvasive imaging system called Raman spectroscopy. Named after Indian scientist C.V. Raman, the technique takes advantage of the fact that certain molecular bonds vibrate at specific frequencies. When a monochromatic laser illuminates a molecular sample, the molecules scatter the light back in various ways depending on their natural vibrations.
However, Xie says, the signal from Raman spectroscopy is weak, particularly when applied to living tissue, where molecular composition is heterogeneous. A specific molecular signal could be lost amidst other backscattered noise. To improve sensitivity, Xie and graduate students Brian Saar and Christian Freudiger developed a high-speed imaging setup with two lasers instead of the conventional one, exploiting a process known as stimulated Raman spectroscopy. The scientists’ goal is to produce label-free images of a wide range of molecules in living animals and humans.
Using knowledge of the vibrational frequency of a specific protein in red blood cells, the team set one laser beam at a high frequency and the other at a lower frequency so that the difference between the two frequencies equaled the vibrational frequency of the protein. Through a system of mirrors, they trained the two laser beams through a small aperture and onto a mouse. The combination of the two laser beams excited the protein molecules in the imaging area and caused them all to vibrate, or oscillate, in synchrony. Xie likens the phenomenon to resetting a group of pendulums.
“If you have lots of pendulums, each one oscillates at the same frequency, but they are randomly distributed in their phase,” says Xie. “That’s what happens in conventional Raman spectroscopy. But here, we forced all the pendulums to go left and right at the same time, so in the end we have a much stronger signal.”
The group developed a custom detector to pick up the molecular signal, and found that the signal was thousands of times stronger than in conventional Raman spectroscopy. From the signal, the researchers obtained fast, detailed images of red blood cells moving through a mouse’s skin capillaries. Using the same technology, they also observed the behavior of trans-retinol, a common skin-care compound, as it was absorbed into a mouse’s ear.
“This work should open new opportunities in studying chemical composition changes and drug transport,” says Shuming Nie, professor of biomedical engineering at Emory University. “This technique is dramatically more sensitive and [has] better spatial resolution, but it is still limited by very small penetration depths.”
So far, Xie and his colleagues have only been able to image at a depth of 100 microns, mostly due to the limitations of laser-based techniques. Xie says a solution may be to pair the technique with an imaging system like MRI, which can produce images deeper in the body, though with less clarity.
The team is working with mechanical engineer Eric Seibel at the University of Washington to design an endoscope that can house the two-laser system, in order to thread it through the body and create detailed images of tissues and organs. With such a capability, says Ji-Xin Cheng, associate professor of biomedical engineering at Purdue University, doctors may be able to identify other diseases that manifest on the surface of organs other than skin.
“Some cancers start in the epithelial layer, or the surface of tissues, like colon cancer,” says Cheng. “Diagnosing such cancers could be a good application for a system like this.”
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.