Why you should care
The ability to diagnose disease is becoming faster and more accurate. Will this lead to greater peace of mind?
You feel a horrible stomachache coming on. Are you imagining it? You didn’t eat anything unusual and you don’t feel sick. What could be causing the pain, and how can you make it go away? Now imagine having a medical device that can quickly read your sweat and tell you whether your symptoms are serious and how to best treat them.
Biopsy results can take seven to 10 days, and even simple STD tests take at least 24 hours. Waiting to get a diagnosis can be incredibly stressful. Now, a growing set of researchers and physicians are developing more accurate, quicker medical diagnostics than ever before, paving the way to a future where we may soon be able to test our symptoms in real time and avoid what we all know as hypochondria.
Scientists at University College London Hospital (UCLH) are working with Google’s DeepMind artificial intelligence to design a system that will analyze brain scans faster than doctors can. The technology has the proven ability to accelerate head and neck cancer diagnoses. Genomic Health in California has developed a genomic test called Oncotype DX that enables oncologists to predict how breast cancer patients will benefit from adjuvant chemotherapy, avoiding unnecessary treatment. Launched in 2004, the test has now been used by nearly 1 million early-stage breast cancer patients in more than 90 countries, giving their physicians the ability to personalize their treatment plans.
If you’re starting to get a cold, or you’re starting to get asthma, it will [intervene] before it becomes chronic.
Antonio Tricoli, Australian National University
In 2014, Kaiser Permanente developed a calculator for health care professionals that can predict the occurrence of early-onset neonatal sepsis, a bacterial infection that can develop in newborns. The tool has already been used more than 750,000 times and has been shown to reduce antibiotic use in infants by 50 percent. And in June 2018, researchers at the Nanotechnology Research Laboratory of Australian National University (ANU) announced that they had designed small optical sensors that they will build into a wearable device, which will medically diagnose health conditions in real time. The sensors — 50 times smaller than a human hair — are designed to respond to changes in the biomolecules from your breath or your sweat, to identify illnesses and eventually recommend proper treatment.
“Once the wearable sensor can measure important biomarkers, you [will] see patterns,” says Antonio Tricoli, leader of the ANU team. “If you’re starting to get a cold, or you’re starting to get asthma, it will [intervene] before it becomes chronic.”
Though the WHO dropped the term ”hypochondria” in 2013, the Diagnostic and Statistical Manual of Mental Disorders recognizes the medical condition now as somatic symptom disorder and illness anxiety disorder. According to WHO, between 1 and 5 percent of the global population, or between 77 million and 385 million people, are estimated to suffer from health anxiety in some form. The condition most commonly emerges in people between the ages of 20 and 40.
For those with health anxiety, searching on their own online about their symptoms isn’t always smart. A 2016 study in the respected journal JAMA found that people who checked their symptoms online got the right answer the first time just 34 percent of the time. Doctors, on the other hand, were right 72 percent of the time. Researchers and organizations developing these advanced diagnostic tools agree that accurate health data coming from a medical professional can serve to reduce health anxiety.
The market and the demand for simple and efficient diagnostics has drawn companies of all shades to the industry — including dubious ones. Notoriously, Elizabeth Holmes and her biotech startup Theranos developed a portable device that claimed to diagnose illnesses such as autoimmune diseases and thyroid conditions, among others, using only the blood from a finger prick. In 2015, investigations by The Wall Street Journal revealed that the device did not work as the company alleged, and Holmes now faces criminal fraud charges in a California federal court.
Still, the odd bad apple hasn’t kept other scientists from innovating.
The ANU sensors, for instance, will not need a blood sample to operate, eliminating all invasive procedures. What’s more, they will not require batteries, wires or lab equipment, making them versatile and accessible. “The cost for medical diagnostics is very expensive,” says Tricoli. “But with miniature technology, we can eliminate [some of the cost].”
UCLH and DeepMind began their partnership in 2016. Two years later, the project is well on its way to an AI system that can analyze head and neck scans with accuracy close to that of a clinician, and in a fraction of the time needed by conventional methods, meaning patients are getting the most appropriate treatment much faster.
Similarly, the Oncotype DX test developed by Genomic Health gives oncologists the ability to diagnose the specific type of breast cancer a patient has and predict the likelihood they will benefit from chemotherapy. The test shows whether the breast cancer patient can be treated with hormone therapy alone, sparing them the adverse effects of chemo.
Predicting risk in conjunction with providing diagnosis can be useful too. Kaiser Permanente’s online sepsis risk calculator avoids the use of unnecessary antibiotics for sepsis in newborns — a condition that is rare now. “Before the risk calculator, about 5-10 percent of live births would be treated for sepsis,” says Dr. Michael Kuzniewicz, director of the perinatal research unit at Kaiser Permanente. Now, antibiotic use has been halved. This also eliminates the need to wait for a blood culture, something that can be very disconcerting for new parents, Kuzniewicz says. “It gives both parents and physicians peace of mind.”
Not unlike Tricoli, Lexington, Massachusetts-based biotech company Quanterix is focused on using biomarkers to more accurately diagnose patients. A technology developed by the company in 2013, called Simoa, measures biomarkers in blood to predict and detect disease at a much earlier stage than conventional methods. Simoa is currently being tested in several therapeutic areas, including oncology, neurology, cardiology and infectious disease. Quanterix CEO, Kevin Hrusovsky, is proud of his company’s advancements but is skeptical about how quickly new diagnostic technology can scale. “Funding is one of the biggest obstacles,” Hrusovsky says. Quanterix was able to pull off an IPO in about 10 years, but this is unusual. Interestingly, there is what’s known in the biotech industry as “the Theranos effect.” Investors are much more adamant today that bio and healthcare-focused startups prove the science behind their technology. Quanterix does this by publishing all of their research in peer-reviewed journals, securing third-party validation.
But while Quanterix is diligent in developing their technology, Hrusovsky remains skeptical about the meaning of “real-time” diagnostics. With cancer, for example, there are very high occurrences of false positives, Hrusovsky says. “Sometimes you don’t know right away if the cancer is lethal. Could ‘real-time’ diagnostics create more false positives?” he asks.
That’s something many in the industry are thinking about. Genomic Health, UCLH and DeepMind are tailoring their technology to ensure that the right treatment plan is part of the diagnostics process. Tricoli agrees that the efficacy of treatment should be a priority. “That’s the final aspect in this,” he says. After receiving treatment, you should “know within weeks if it worked,” he adds.
Once enough of these technologies do work, hypochondriacs may need to find something else to worry about.
Corrections: The original version of this feature stated the wrong year for the development of the risk calculator, and mistook genomic for genetic in one instance.