In an initial study, Canary analyzed speech recordings with AI and identified early stage Alzheimer’s with 96 percent accuracy.
Dave Arnold retired in his 60s and began spending time volunteering in local schools. But then he started misplacing items, forgetting appointments and losing his sense of direction. Eventually he was diagnosed with early stage Alzheimer’s.
“Hearing the diagnosis made me very emotional and tearful,” he said. “I immediately thought of all my mom had experienced.” His mother suffered with the condition for years before passing away. Over the last year, Arnold has worked for the Alzheimer’s Association as one of its early stage advisors, sharing his insights to help others in the initial stages of the disease.
Arnold was diagnosed sooner than many others. It's important to find out early, when interventions can make the most difference. One promising avenue is looking at how people talk. Research has shown that Alzheimer’s affects a part of the brain that controls speech, resulting in small changes before people show other signs of the disease.
Now, Canary Speech, a company based in Utah, is using AI to examine elements like the pitch of a person’s voice and their pauses. In an initial study, Canary analyzed speech recordings with AI and identified early stage Alzheimer’s with 96 percent accuracy.
Developing the AI model
Canary Speech’s CEO, Henry O’Connell, met cofounder Jeff Adams about 40 years before they started the company. Back when they first crossed paths, they were both living in Bethesda, Maryland; O’Connell was a research fellow at the National Institutes of Health studying rare neurological diseases, while Adams was working to decode spy messages. Later on, Adams would specialize in building mathematical models to analyze speech and sound as a team leader in developing Amazon's Alexa.
It wasn't until 2015 that they decided to make use of the fit between their backgrounds. ““We established Canary Speech in 2017 to build a product that could be used in multiple languages in clinical environments,” O'Connell says.
The need is growing. About 55 million people worldwide currently live with Alzheimer’s, a number that is expected to double by 2050. Some scientists think the disease results from a buildup of plaque in the brain. It causes mild memory loss at first and, over time, this issue get worse while other symptoms, such as disorientation and hallucinations, can develop. Treatment to manage the disease is more effective in the earlier stages, but detection is difficult since mild symptoms are often attributed to the normal aging process.
O’Connell and Adams specialize in the complex ways that Alzheimer’s effects how people speak. Using AI, their mathematical model analyzes 15 million data points every minute, focusing on certain features of speech such as pitch, pauses and elongation of words. It also pays attention to how the vibrations of vocal cords change in different stages of the disease.
To create their model, the team used a type of machine learning called deep neural nets, which looks at multiple layers of data - in this case, the multiple features of a person’s speech patterns.
“Deep neural nets allow us to look at much, much larger data sets built out of millions of elements,” O’Connell explained. “Through machine learning and AI, we’ve identified features that are very sensitive to an Alzheimer’s patient versus [people without the disease] and also very sensitive to mild cognitive impairment, early stage and moderate Alzheimer's.” Based on their learnings, Canary is able to classify the disease stage very quickly, O’Connell said.
“When we’re listening to sublanguage elements, we’re really analyzing the direct result of changes in the brain in the physical body,” O’Connell said. “The brain controls your vocal cords: how fast they vibrate, the expansion of them, the contraction.” These factors, along with where people put their tongues when talking, function subconsciously and result in subtle changes in the sounds of speech.
Further testing is needed
In an initial trial, Canary analyzed speech recordings from phone calls to a large U.S. health insurer. They looked at the audio recordings of 651 policyholders who had early stage Alzheimer’s and 1018 who did not have the condition, aiming for a representative sample of age, gender and race. They used this data to create their first diagnostic model and found that it was 96 percent accurate in identifying Alzheimer’s.
Christian Herff, an assistant professor of neuroscience at Maastricht University in the Netherlands, praised this approach while adding that further testing is needed to assess its effectiveness.
“I think the general idea of identifying increased risk for cognitive impairment based on speech characteristics is very feasible, particularly when change in a user’s voice is monitored, for example, by recording speech every year,” Herff said. He noted that this can only be a first indication, not a full diagnosis. The accuracy still needs to be validated in studies that follows individuals over a period of time, he said.
Toby Walsh, a professor of artificial intelligence at the University of New South Wales, also thinks Canary’s tool has potential but highlights that Canary could diagnose some people who don’t really have the disease. “This is an interesting and promising application of AI,” he said, “but these tools need to be used carefully. Imagine the anxiety of being misdiagnosed with Alzheimer’s.”
As with many other AI tools, privacy and bias are additional issues to monitor closely, Walsh said.
Other languages
A related issue is that not everyone is fluent in English. Mahnaz Arvaneh, a senior lecturer in automatic control and systems engineering at the University of Sheffield, said this could be a blind spot.
“The system may not be very accurate for those who have English as their second language as their speaking patterns would be different, and any issue might be because of language deficiency rather than cognitive issues,” Arvaneh said.
The team is expanding to multiple languages starting with Japanese and Spanish. The elements of the model that make up the algorithm are very similar, but they need to be validated and retrained in a different language, which will require access to more data.
Recently, Canary analyzed the phone calls of 233 Japanese patients who had mild cognitive impairment and 704 healthy people. Using an English model they were able to identify the Japanese patients who had mild cognitive impairment with 78 percent accuracy. They also developed a model in Japanese that was 45 percent accurate, and they’re continuing to train it with more data.
The future
Canary is using their model to look at other diseases like Huntington’s and Parkinson’s. They’re also collaborating with pharmaceuticals to validate potential therapies for Alzheimer’s. By looking at speech patterns over time, Canary can get an indication of how well these drugs are working.
Dave Arnold and his wife dance at his nephew’s wedding in Rochester, New York, ten years ago, before his Alzheimer's diagnosis.
Dave Arnold
Ultimately, they want to integrate their tool into everyday life. “We want it to be used in a smartphone, or a teleconference call so that individuals could be examined in their home,” O’Connell said. “We could follow them over time and work with clinical teams and hospitals to improve the evaluation of patients and contribute towards an accurate diagnosis.”
Arnold, the patient with early stage Alzheimer’s, sees great promise. “The process of getting a diagnosis is already filled with so much anxiety,” he said. “Anything that can be done to make it easier and less stressful would be a good thing, as long as it’s proven accurate.”
Doctors worry that fungal pathogens may cause the next pandemic.
Although C. auris typically doesn’t sicken healthy people, it afflicts immunocompromised hospital patients and may cause severe infections that can lead to sepsis, a life-threatening condition in which the overwhelmed immune system begins to attack the body’s own organs. Between 30 and 60 percent of patients who contract a C. auris infection die from it, according to the CDC. People who are undergoing stem cell transplants, have catheters or have taken antifungal or antibiotic medicines are at highest risk. “We’re coming to a perfect storm of increasing resistance rates, increasing numbers of immunosuppressed patients worldwide and a bug that is adapting to higher temperatures as the climate changes,” says Prabhavathi Fernandes, chair of the National BioDefense Science Board.
Most Candida species aren’t well-adapted to our body temperatures so they aren’t a threat. C. auris, however, thrives at human body temperatures.
Although medical professionals aren’t concerned at this point about C. auris evolving to affect healthy people, they worry that its presence in hospitals can turn routine surgeries into life-threatening calamities. “It’s coming,” says Fernandes. “It’s just a matter of time.”
An emerging global threat
“Fungi are found in the environment,” explains Fernandes, so Candida spores can easily wind up on people’s skin. In hospitals, they can be transferred from contact with healthcare workers or contaminated surfaces. Most Candida species aren’t well-adapted to our body temperatures so they aren’t a threat. C. auris, however, thrives at human body temperatures. It can enter the body during medical treatments that break the skin—and cause an infection. Overall, fungal infections cost some $48 billion in the U.S. each year. And infection rates are increasing because, in an ironic twist, advanced medical therapies are enabling severely ill patients to live longer and, therefore, be exposed to this pathogen.
The first-ever case of a C. auris infection was reported in Japan in 2009, although an analysis of Candida samples dated the earliest strain to a 1996 sample from South Korea. Since then, five separate varieties – called clades, which are similar to strains among bacteria – developed independently in different geographies: South Asia, East Asia, South Africa, South America and, recently, Iran. So far, C. auris infections have been reported in 35 countries.
In the U.S., the first infection was reported in 2016, and the CDC started tracking it nationally two years later. During that time, 5,654 cases have been reported to the CDC, which only tracks U.S. data.
What’s more notable than the number of cases is their rate of increase. In 2016, new cases increased by 175 percent and, on average, they have approximately doubled every year. From 2016 through 2022, the number of infections jumped from 63 to 2,377, a roughly 37-fold increase.
“This reminds me of what we saw with epidemics from 2013 through 2020… with Ebola, Zika and the COVID-19 pandemic,” says Robin Robinson, CEO of Spriovas and founding director of the Biomedical Advanced Research and Development Authority (BARDA), which is part of the U.S. Department of Health and Human Services. These epidemics started with a hockey stick trajectory, Robinson says—a gradual growth leading to a sharp spike, just like the shape of a hockey stick.
Another challenge is that right now medics don’t have rapid diagnostic tests for fungal infections. Currently, patients are often misdiagnosed because C. auris resembles several other easily treated fungi. Or they are diagnosed long after the infection begins and is harder to treat.
The problem is that existing diagnostics tests can only identify C. auris once it reaches the bloodstream. Yet, because this pathogen infects bodily tissues first, it should be possible to catch it much earlier before it becomes life-threatening. “We have to diagnose it before it reaches the bloodstream,” Walsh says.
The most alarming fact is that some Candida infections no longer respond to standard therapeutics.
“We need to focus on rapid diagnostic tests that do not rely on a positive blood culture,” says John Sperzel, president and CEO of T2 Biosystems, a company specializing in diagnostics solutions. Blood cultures typically take two to three days for the concentration of Candida to become large enough to detect. The company’s novel test detects about 90 percent of Candida species within three to five hours—thanks to its ability to spot minute quantities of the pathogen in blood samples instead of waiting for them to incubate and proliferate.
Unlike other Candida species C. auris thrives at human body temperatures
Adobe Stock
Tackling the resistance challenge
The most alarming fact is that some Candida infections no longer respond to standard therapeutics. The number of cases that stopped responding to echinocandin, the first-line therapy for most Candida infections, tripled in 2020, according to a study by the CDC.
Now, each of the first four clades shows varying levels of resistance to all three commonly prescribed classes of antifungal medications, such as azoles, echinocandins, and polyenes. For example, 97 percent of infections from C. auris Clade I are resistant to fluconazole, 54 percent to voriconazole and 30 percent of amphotericin. Nearly half are resistant to multiple antifungal drugs. Even with Clade II fungi, which has the least resistance of all the clades, 11 to 14 percent have become resistant to fluconazole.
Anti-fungal therapies typically target specific chemical compounds present on fungi’s cell membranes, but not on human cells—otherwise the medicine would cause damage to our own tissues. Fluconazole and other azole antifungals target a compound called ergosterol, preventing the fungal cells from replicating. Over the years, however, C. auris evolved to resist it, so existing fungal medications don’t work as well anymore.
A newer class of drugs called echinocandins targets a different part of the fungal cell. “The echinocandins – like caspofungin – inhibit (a part of the fungi) involved in making glucan, which is an essential component of the fungal cell wall and is not found in human cells,” Fernandes says. New antifungal treatments are needed, she adds, but there are only a few magic bullets that will hit just the fungus and not the human cells.
Research to fight infections also has been challenged by a lack of government support. That is changing now that BARDA is requesting proposals to develop novel antifungals. “The scope includes C. auris, as well as antifungals following a radiological/nuclear emergency, says BARDA spokesperson Elleen Kane.
The remaining challenge is the number of patients available to participate in clinical trials. Large numbers are needed, but the available patients are quite sick and often die before trials can be completed. Consequently, few biopharmaceutical companies are developing new treatments for C. auris.
ClinicalTrials.gov reports only two drugs in development for invasive C. auris infections—those than can spread throughout the body rather than localize in one particular area, like throat or vaginal infections: ibrexafungerp by Scynexis, Inc., fosmanogepix, by Pfizer.
Scynexis’ ibrexafungerp appears active against C. auris and other emerging, drug-resistant pathogens. The FDA recently approved it as a therapy for vaginal yeast infections and it is undergoing Phase III clinical trials against invasive candidiasis in an attempt to keep the infection from spreading.
“Ibreafungerp is structurally different from other echinocandins,” Fernandes says, because it targets a different part of the fungus. “We’re lucky it has activity against C. auris.”
Pfizer’s fosmanogepix is in Phase II clinical trials for patients with invasive fungal infections caused by multiple Candida species. Results are showing significantly better survival rates for people taking fosmanogepix.
Although C. auris does pose a serious threat to healthcare worldwide, scientists try to stay optimistic—because they recognized the problem early enough, they might have solutions in place before the perfect storm hits. “There is a bit of hope,” says Robinson. “BARDA has finally been able to fund the development of new antifungal agents and, hopefully, this year we can get several new classes of antifungals into development.”