A physician is faced with 3 patients: an elderly person with a chronic cough, a child being wheeled out of surgery and a young mother with a high fever. How does the doctor identify which of these patients needs antibiotics? The answer: mostly guesswork. And, more than likely, all 3 will be prescribed with a course of antibiotics.
The problem, as described by physician turned entrepreneur Tim Sweeney, lies in the fact that we don’t have a way of accurately detecting the early stages of bacterial or viral infections. While hospitals routinely use blood tests, the majority of pathogenic infections are, in fact, not detectable in the blood circulation. To play it safe, physicians often administer antibiotics as a precautionary measure, especially considering the drug-free alternative could mean an elevated risk of sepsis.
During sepsis, an extreme inflammatory response floods the body with chemicals to fight the infection. This process can snowball, triggering dangerous and potentially life-threatening chemical imbalances, culminating in irreversible damage to multiple organ systems. Sepsis represents a major clinical crisis, responsible for half of all deaths in hospital and is the most expensive diagnosis in the US healthcare system.
The unconstrained use of antibiotics, however, is not the answer. Not only do these drugs pose significant side effects to patients, but they also contribute to a serious global epidemic: the steady rise of the antibiotic-resistant superbug.
Sweeney, together with a team of computational immunologists at Stanford University envisioned a future where doctors had access to robust diagnostic tools to check for the early signs of infections, leaving the guesswork out of the equation.
Using machine learning methods, they analyzed the genetic activation signatures of multiple cohorts of patients around the world. This technique, called transcriptomic analysis, enabled them to draw out definitive patterns of activation in genes that control inflammation, in patients with bacterial and viral infections.
Fascinatingly, the team discovered specific genetic motifs unique to patients at risk of sepsis. Analyzing a cluster of just 7 genes gave the researchers enough information on the presence and severity of the infection, even highlighting whether it is bacterial or viral in nature.
Expanding this to a set of 30 genes increased the granularity of the diagnosis, including what type of treatment would be most effective, downstream diagnostic tests to run and whether or not the individual needs urgent emergency care. All this in under 30 minutes.
This innovation has been commercialized by the start-up company Inflammatix, of which Sweeny is a co-founder. Inflammatix has been creating a buzz recently after getting a $6 million funding injection from the Biomedical Advanced Research and Development Authority (BARDA) to further develop its platform for diagnosing both sepsis and influenza.
If all goes well, the deal has the potential to be worth up to $72 million - a relatively small price for the lifesaving potential of this technology.