Claire Guest, co-founder of Medical Detection Dogs, with Daisy, whom she credits with saving her life.
Daisy wouldn't leave Claire Guest alone. Instead of joining Guest's other dogs for a run in the park, the golden retriever with the soulful eyes kept nudging Guest's chest, and stared at her intently, somehow hoping she'd get the message.
"I was incredibly lucky to be told by Daisy."
When Guest got home, she detected a tiny lump in one of her breasts. She dismissed it, but her sister, who is a family doctor, insisted she get it checked out.
That saved her life. A series of tests, including a biopsy and a mammogram, revealed the cyst was benign. But doctors discovered a tumor hidden deep inside her chest wall, an insidious malignancy that normally isn't detected until the cancer has rampaged out of control throughout the body. "My prognosis would have been very poor," says Guest, who is an animal behavioralist. "I was incredibly lucky to be told by Daisy."
Ironically, at the time, Guest was training hearing dogs for the deaf—alerting them to doorbells or phones--for a charitable foundation. But she had been working on a side project to harness dogs' exquisitely sensitive sense of smell to spot cancer at its earliest and most treatable stages. When Guest was diagnosed with cancer two decades ago, however, the use of dogs to detect diseases was in its infancy and scientific evidence was largely anecdotal.
In the years since, Guest and the British charitable foundation she co-founded with Dr. John Church in 2008, Medical Detection Dogs (MDD), has shown that dogs can be trained to detect odors that predict a looming medical crisis hours in advance, in the case of diabetes or epilepsy, as well as the presence of cancers.
In a proof of principle study published in the BMJ in 2004, they showed dogs had better than a 40 percent success rate in identifying bladder cancer, which was significantly better than random chance (14 percent). Subsequent research indicated dogs can detect odors down to parts per trillion, which is the equivalent of sniffing out a teaspoon of sugar in two Olympic size swimming pools (a million gallons).
American scientists are devising artificial noses that mimic dogs' sense of smell, so these potentially life-saving diagnostic tools are widely available.
But the problem is "dogs can't be scaled up"—it costs upwards of $25,000 to train them—"and you can't keep a trained dog in every oncology practice," says Guest.
The good news is that the pivotal 2004 BMJ paper caught the attention of two American scientists—Andreas Mershin, a physicist at MIT, and Wen-Yee Yee, a chemistry professor at The University of Texas at El Paso. They have joined Guest's quest to leverage canines' highly attuned olfactory systems and devise artificial noses that mimic dogs' sense of smell, so these potentially life-saving diagnostic tools are widely available.
"What we do know is that this is real," says Guest. "Anything that can improve diagnosis of cancer is something we ought to know about."
Dogs have routinely been used for centuries as trackers for hunting and more recently, for ferreting out bombs and bodies. Dogs like Daisy, who went on to become a star performer in Guest's pack of highly trained cancer detecting canines before her death in 2018, have shared a special bond with their human companions for thousands of years. But their vastly superior olfaction is the result of simple anatomy.
Humans possess about six million olfactory receptors—the antenna-like structures inside cell membranes in our nose that latch on to the molecules in the air when we inhale. In contrast, dogs have about 300 million of them and the brain region that analyzes smells is, proportionally, about 40 times greater than ours.
Research indicates that cancerous cells interfere with normal metabolic processes, prompting them to produce volatile organic compounds (VOCs), which enter the blood stream and are either exhaled in our breath or excreted in urine. Dogs can identify these VOCs in urine samples at the tiniest concentrations, 0.001 parts per million, and can be trained to identify the specific "odor fingerprint" of different cancers, although teaching them how to distinguish these signals from background odors is far more complicated than training them to detect drugs or explosives.
For the past fifteen years, Andreas Mershin of MIT has been grappling with this complexity in his quest to devise an artificial nose, which he calls the Nano-Nose, first as a military tool to spot land mines and IEDS, and more recently as a cancer detection tool that can be used in doctors' offices. The ultimate goal is to create an easy-to-use olfaction system powered by artificial intelligence that can fit inside of smartphones and can replicate dogs' ability to sniff out early signs of prostate cancer, which could eliminate a lot of painful and costly biopsies.
Trained canines have a better than 90 percent accuracy in spotting prostate cancer, which is normally difficult to detect. The current diagnostic, the prostate specific antigen test, which measures levels of certain immune system cells associated with prostate cancer, has about as much accuracy "as a coin toss," according to the scientist who discovered PSA. These false positives can lead to unnecessary and horrifically invasive biopsies to retrieve tissue samples.
So far, Mershin's prototype device has the same sensitivity as the dogs—and can detect odors at parts per trillion—but it still can't distinguish that cancer smell in individual human patients the way a dog can. "What we're trying to understand from the dogs is how they look at the data they are collecting so we can copy it," says Mershin. "We still have to make it intelligent enough to know what it is looking at—what we are lacking is artificial dog intelligence."
The intricate parts of the artificial nose are designed to fit inside a smartphone.
At UT El Paso, Wen-Yee Lee and her research team has used the canine olfactory system as a model for a new screening test for prostate cancer, which has a 92 percent accuracy in tests of urine samples and could be eventually developed as a kit similar to the home pregnancy test. "If dogs can do it, we can do it better," says Lee, whose husband was diagnosed with prostate cancer in 2005.
The UT scientists used samples from about 150 patients, and looked at about 9,000 compounds before they were able to zero in on the key VOCs that are released by prostate cancers—"it was like finding a needle in the haystack," says Lee. But a more reliable test that can also distinguish which cancers are more aggressive could help patients decide their best treatment options and avoid invasive procedures that can render them incontinent and impotent.
"This is much more accurate than the PSA—we were able to see a very distinct difference between people with prostate cancer and those without cancer," says Lee, who has been sharing her research with Guest and hopes to have the test on the market within the next few years.
In the meantime, Guest's foundation has drawn the approving attention of royal animal lovers: Camilla, the Duchess of Cornwall, is a patron, which opened up the charitable floodgates and helped legitimize MDD in the scientific community. Even Camilla's mother-in-law, Queen Elizabeth, has had a demonstration of these canny canines' unique abilities.
Claire Guest, and two of MDDs medical detection dogs, Jodie and Nimbus, meet with queen Elizabeth.
"She actually held one of my [artificial] noses in her hand and asked really good questions, including things we hadn't thought of, like the range of how far away a dog can pick up the scent or if this can be used to screen for malaria," says Mershin. "I was floored by this curious 93-year-old lady. Half of humanity's deaths are from chronic diseases and what the dogs are showing is a whole new way of understanding holistic diseases of the system."
AI was integral to creating Moderna's mRNA vaccine against COVID.
But the truth is that before ChatGPT nabbed the public’s attention, AI was already here, and it was doing more important things than writing essays for lazy college students. Case in point: It was key to saving the lives of tens of millions of people.
AI-designed mRNA vaccines
As Dave Johnson, chief data and AI officer at Moderna, told MIT Technology Review‘s In Machines We Trust podcast in 2022, AI was integral to creating the company’s highly effective mRNA vaccine against COVID. Moderna and Pfizer/BioNTech’s mRNA vaccines collectively saved between 15 and 20 million lives, according to one estimate from 2022.
Johnson described how AI was hard at work at Moderna, well before COVID arose to infect billions. The pharmaceutical company focuses on finding mRNA therapies to fight off infectious disease, treat cancer, or thwart genetic illness, among other medical applications. Messenger RNA molecules are essentially molecular instructions for cells that tell them how to create specific proteins, which do everything from fighting infection, to catalyzing reactions, to relaying cellular messages.
Johnson and his team put AI and automated robots to work making lots of different mRNAs for scientists to experiment with. Moderna quickly went from making about 30 per month to more than one thousand. They then created AI algorithms to optimize mRNA to maximize protein production in the body — more bang for the biological buck.
For Johnson and his team’s next trick, they used AI to automate science, itself. Once Moderna’s scientists have an mRNA to experiment with, they do pre-clinical tests in the lab. They then pore over reams of data to see which mRNAs could progress to the next stage: animal trials. This process is long, repetitive, and soul-sucking — ill-suited to a creative scientist but great for a mindless AI algorithm. With scientists’ input, models were made to automate this tedious process.
“We don’t think about AI in the context of replacing humans,” says Dave Johnson, chief data and AI officer at Moderna. “We always think about it in terms of this human-machine collaboration, because they’re good at different things. Humans are really good at creativity and flexibility and insight, whereas machines are really good at precision and giving the exact same result every single time and doing it at scale and speed.”
All these AI systems were in put in place over the past decade. Then COVID showed up. So when the genome sequence of the coronavirus was made public in January 2020, Moderna was off to the races pumping out and testing mRNAs that would tell cells how to manufacture the coronavirus’s spike protein so that the body’s immune system would recognize and destroy it. Within 42 days, the company had an mRNA vaccine ready to be tested in humans. It eventually went into hundreds of millions of arms.
Biotech harnesses the power of AI
Moderna is now turning its attention to other ailments that could be solved with mRNA, and the company is continuing to lean on AI. Scientists are still coming to Johnson with automation requests, which he happily obliges.
“We don’t think about AI in the context of replacing humans,” he told the Me, Myself, and AI podcast. “We always think about it in terms of this human-machine collaboration, because they’re good at different things. Humans are really good at creativity and flexibility and insight, whereas machines are really good at precision and giving the exact same result every single time and doing it at scale and speed.”
Moderna, which was founded as a “digital biotech,” is undoubtedly the poster child of AI use in mRNA vaccines. Moderna recently signed a deal with IBM to use the company’s quantum computers as well as its proprietary generative AI, MoLFormer.
Moderna’s success is encouraging other companies to follow its example. In January, BioNTech, which partnered with Pfizer to make the other highly effective mRNA vaccine against COVID, acquired the company InstaDeep for $440 million to implement its machine learning AI across its mRNA medicine platform. And in May, Chinese technology giant Baidu announced an AI tool that designs super-optimized mRNA sequences in minutes. A nearly countless number of mRNA molecules can code for the same protein, but some are more stable and result in the production of more proteins. Baidu’s AI, called “LinearDesign,” finds these mRNAs. The company licensed the tool to French pharmaceutical company Sanofi.
Writing in the journal Accounts of Chemical Research in late 2021, Sebastian M. Castillo-Hair and Georg Seelig, computer engineers who focus on synthetic biology at the University of Washington, forecast that AI machine learning models will further accelerate the biotechnology research process, putting mRNA medicine into overdrive to the benefit of all.
This article originally appeared on Big Think, home of the brightest minds and biggest ideas of all time.
