For many of the most severe health conditions, early detection greatly increases the chance of successful treatment and much better outcomes. This article shows how some surprising technologies are improving the early detection of several conditions. Nutromics continuously monitors the delivery of an antibiotic used to treat sepsis. Eye-tracking from Tobii is used by several of its clients to detect certain kinds of autism as early as twelve months of age. And Cordio Medical is detecting oncoming congestive heart failure (CHF) three weeks in advance of an incident.
Minimally Invasive Monitoring Provides Real-Time Measurements
Although doctors learn what people have in their bodies most often through blood tests, technologies developed over the past few decades can find a lot of information without inserting a large needle and removing blood from veins. According to Dr. Agim Beshiri, chief medical officer of Nutromics, a lot of molecules are accessible near the skin’s surface through interstitial fluid (ISF) and even sweat. Nutromics takes advantage of this with an innovative, minimally invasive, methodology to check for various conditions.
First, they research the molecule that they’re trying to detect. They then create a synthetic DNA strand that can latch onto it, in the same manner that pharma companies use mRNA or DNA to develop vaccines and deliver drugs into the body.
Nutromics developed a “lab-on-a-patch,” about two centimeters or one inch long (Figure 1), that contains a few tiny needles. These needles penetrate only one millimeter into the patient’s skin—even less in infants. By comparison, the continuous glucose monitors that have been used for the past two decades penetrate about ten millimeters.
When the DNA on the end of a needle latches onto a molecule, the DNA bends, creating an electric signal that the device detects. The DNA releases the molecule, so that the device can continue monitoring. This method was actually developed to mimic how biological sensors in our body work around the clock. By measuring the rate of the electronic signals, a Nutromics app can determine the amount of the detected molecule in the patient’s body. A one-minute video shows more about how the system works.
The first application for Nutromics, still in the trial phase, measures the antibiotic vancomycin, which is used to treat severe infections like sepsis and methicillin-resistant staphylococcus aureus (MRSA) infections. According to Peter Vranes, CEO and co-founder of Nutromics, an improved method of monitoring vancomycin is critical, because the current method is not personalized and has room for improvement.
Currently, when hospitals monitor a patient being treated with vancomycin, they take a blood test that requires several hours to return results. Although the treatment regimen calls for testing once or twice a day at very specific times, many hospitals test less often due to workflow and resource constraints. Consequently, when it’s time to adjust the vancomycin dosage, the doctor is dealing with outdated information.
The difficulty in personalizing doses often results in overdosing or underdosing that can lead to kidney toxicity and even death. Underdosing might not effectively address the infection and can lead to worse outcomes, possibly death. This ends up costing healthcare $13 billion a year. Nutromics is developing a personalized solution that will give doctors continuous, real-time data to resolve the current limitations in healthcare.
Nutromics’s measurement solution will be available commercially in 2028. They plan many other applications for the Nutromics patch.
Autism Detection Through Eye Tracking
Eye tracking is fairly standard technology these days. Tobii lists attention, memory, perception, language, problem-solving, and decision-making among the personal attributes it can measure through eye-tracking, particularly a technology called pupil center corneal reflection (PCCR).
Eye-tracking has been the focus of Dr. Karen Pierce for a decade, especially in the detection of autism. In a half-hour video, she describes how certain types of autism have a high rate of detection between the ages of one and two through a ten-minute computer evaluation using Tobii eye-tracking (Figure 2).
Basically, the app tracks the baby’s reactions to faces, and their preference for a non-social picture versus a social one. The diagnosis delivered by the computer app matches closely to brain activity and behaviors associated with autistic people.
This diagnostic tool is a big advance, and won FDA approval in 2022. Pierce says the test can also guide the prognosis for a patient.
Marisa Biondi, Senior Research Scientist at Tobii, tells me autism is traditionally diagnosed through an Autism Diagnostic Observation Schedule (ADOS) that requires a multi-hour interview observation of the patient by a highly trained professional, who combines their observations with information provided by parents. This test is usually not done until later in the second year of life, and can be quite subjective. Pierce’s app is more objective and faster, and allows intervention to start much earlier, which is critical for language development and other social skills.
More background on the technology can be found in a research article, Eye Tracking Young Children with Autism, and a webinar by Pierce.
CHF Detection Through Vocal Markers
The analysis of an individual’s voice, which can be done by an app loaded on their mobile phone, is in common use to measure depression, anxiety, dementia, and other conditions related to emotional and mental functioning. But how can a vocal marker reveal congestive heart failure?
The key link, according to Cordio Medical CEO Tamir Tal, is the presence of fluid in the patients’ lungs. As the lungs fill up, the voice changes. When the fluid reaches a certain point, it leads to heart failure.
The Cordio Medical app that checks the presence of fluid in the lungs is called HearO. It has to start with a patient in remission, so the app was used first upon hospital discharge. The app was developed to allow doctors to monitor a patient’s condition remotely on a daily basis.
For seven to ten days, the app takes samples of the patient’s healthy voice. So long as the patient remains healthy, the baseline is updated regularly with new vocal samples. The app compares these samples to later samples and detect a change suggesting that the lungs are filling with fluid.
According to a video by Dr. William T. Abraham of Ohio State University, a study of 253 patients and 650,00 voice recordings showed that HearO is much more effective than traditional diagnostic tools.
If no diagnostic tools are available, the patient or doctor is warned about heart failure by wheezing and shortness of breath. But these occur very shortly before a crisis. More sophisticated techniques measure the patient’s weight or oxygen saturation, but these can predict heart failure only 35% of the time.
In contrast, the HearO vocal marker technique predicts a heart failure 80% of the time, with very few false positives, 18 days prior to the event on average. This can save lives. Having completed their clinical study, Cordio Medical is submitting the app for FDA approval.
I should note that vocal marker analysis is getting so powerful that it’s scary. An article titled Your Voice Gave You Away: The Privacy Risks of Voice-Inferred Information discusses the risks. But when freely chosen by an individual, and seeded on purpose with that individual’s baseline, the technology is an important tool in medicine.
