Even when an infection is diagnosed, the treatment plan isn’t always clear. Some sepsis patients, for instance, recover well with steroid treatment, while others react poorly and their condition declines. But clues in a patient’s immune system response could help physicians quickly and accurately zero in on a plan of action.
In two recent scientific papers, Purvesh Khatri, PhD, professor of biomedical informatics, has laid out a road map for such a tool, detailing the development and validation of a collection of blood tests that would provide answers to all of these questions. Those tests could readily assist emergency clinicians in determining diagnosis and care protocols for patients with a suspected infection or critical condition — those experiencing sepsis, burn trauma, infection or acute respiratory distress.
Together, the tests — which assess genetic activity patterns, also known as signatures, in immune cells to reveal a patient’s immune state — could help physicians assess whether patients need treatment and, if so, the most beneficial type.
Past studies conducted by Khatri and his team have shown that immune cell gene signatures can diagnose the existence of an infection as well as predict type and severity. The team used that information to translate certain gene signatures into a test useable in the clinic, which received U.S. Food and Drug Administration clearance earlier this year.
It made Khatri wonder if blood-based signatures could also guide treatment decisions for severe infections and other critical conditions frequently seen in intensive care units. To answer this question, the team developed a scoring system that quantified immune cell dysregulation by identifying “good” gene signatures, which indicated a desired healthy immune response, and “bad” ones, which signified an imbalanced, detrimental immune response associated with a higher risk of severe outcomes and a need for prompt treatment.
“This work, combined with the fact that we have an FDA-cleared clinical test, is an indicator that we are likely at the beginning of the era of precision medicine in critical care,” said Khatri, who is a member of the Stanford Institute for Immunity, Transplantation and Infection. “We finally have all the required tools to match the right people with the right treatment at the right time.”
Two papers describing the studies will be published in Nature Medicine on Sept. 30. Khatri; Angela Rogers, MD, associate professor of pulmonary and critical care; and Timothy Sweeney, MD, PhD, a former Stanford postdoc, are co-senior authors of the paper that describes the treatment-focused study. Andrew Moore, MD, postdoc in the Institute for Immunity, Transplantation and Infection, and a clinical instructor of pulmonary, allergy and critical care medicine, is the first author of that study. Oliver Liesenfeld, MD, former chief medical officer and current advisor at Inflammatix is the lead and corresponding author of the study that validated the use of the diagnostic tests in a clinical setting. Sweeney, Khatri and Nathan Shapiro, MD, professor of emergency medicine at Harvard Medical School, are co-senior authors of that study.
Scoring the immune system
The existing, FDA-approved diagnostic and predictive gene signature test, known as TriVerity, measures the activity of 29 genes and uses artificial intelligence to provide three scores for likelihood of a bacterial infection, a viral infection and severe illness, defined as requiring ICU-level care within seven days.
The recently published TriVerity validation study enrolled 1,222 patients from 22 emergency departments in the United States and Europe, analyzing how well the test identified infection and predicted severity in real health care settings. Overall, the test outperformed three of the clinical standards used to flag infection, providing a more accurate diagnosis, prediction of illness severity and better guidance on whether to use antibiotics.
In a separate study that investigated gene signatures to inform critical care treatment, the researchers analyzed more than 7,000 blood samples from 37 cohorts in 13 countries to better understand how immune cell activity might predict severity of a critical illness and response to treatment. The team collected data on critical care cases, treatments and outcomes from public repositories and from a newly formed consortium of 11 institutions. From this information, they proposed a new scoring system, called the Human Immune Dysregulation Evaluation Framework (HI-DEF), which provides two scores that indicate the status of a patient’s immune system health — helping predict whether a patient will have a healthy or unhealthy immune response.
The scores divide the patients into four groups: myeloid dysregulation, lymphoid dysregulation, systemwide dysregulation (in which both myeloid and lymphoid cells are off-kilter) and balanced response (in which all immune cells are functioning as expected). The study showed that an increase in either immune dysregulation score, represented with a higher level of “bad” gene signatures, compared with the “good” ones, was linked to poor outcomes in a range of critical illnesses, including sepsis, burn, trauma and acute respiratory distress.
For most critical illnesses, standard treatments don’t specifically target the immune response. Instead, doctors focus their efforts on treating the symptoms of the infection — for example, administering medications that address their blood pressure or placing patients with failing lungs on a breathing machine.
However, if a patient’s immune system is out of balance, these treatments can backfire. Sepsis, for instance, can stem from viral or bacterial infections, but the symptoms are similar. Bacterial sepsis worsens by the hour and requires a swift prescription of antibiotics — that’s why doctors often immediately administer antibiotics. But it’s not without risk, according to Khatri. If the infection turns out to be viral, the antibiotics are not only ineffective; they can create ideal conditions for antibiotic-resistant bacteria to flourish.
Khatri found that categorizing patients into groups that signify whether and how different arms of their immune system are functioning — or not functioning — can help doctors make better, faster, targeted treatment decisions. For example, if a patient has myeloid dysregulation, they will likely benefit from drugs that target the myeloid immune response, while patients with lymphoid dysregulation need drugs focused on the lymphoid immune response. If both are dysregulated, doctors may choose a combination of lymphoid- and myeloid-targeting drugs. This information can be used to prescribe the proper treatment once, at the beginning, eliminating any guesswork.
Further data analysis also showed that patients with high lymphoid dysregulation, whether they had sepsis or were suffering from burns, generally benefited from steroid treatments; mortality rates improved when appropriately treated. In contrast, when both myeloid and lymphoid immune responses were balanced, patients did not benefit from steroid treatment, and mortality rates worsened. Prospective clinical studies are needed to identify more specific treatments, such as steroids, Khatri said.
Dysregulation beyond the ICU
Khatri and his team plan to pair the HI-DEF scores with the TriVerity diagnostic test to create a one-stop shop that can help doctors analyze blood samples, identify immune system dysregulation and guide treatments in as little as 30 minutes. In the future, Khatri hopes, doctors could take a patient’s blood sample, run it through the tool and receive a gene signature analysis that tells them the patient’s diagnosis as well as whether and how they should be treated.
“You could have a platform to identify the infection, severity of the illness and the treatment quickly,” Khatri said.
Khatri also hopes the dysregulation scoring system will one day move beyond critical illnesses. Signs of immune dysregulation can appear long before a patient ends up in the ICU, so the tool could be used to surveil general health concerns, he said. (Previous work conducted by Khatri found that patients with other high-risk health concerns, such as diabetes, had a higher number of “bad” gene signatures.) More research is needed to determine whether lifestyle changes could alter these signatures, but it’s a signal to Khatri that there’s a strong association warranting more exploration.
“My vision is to make an immune dysregulation assessment part of your annual health checkup,” he said.
Researchers at Inflammatix, Inc.; the University of Amsterdam; the University of Groningen; the University of Malta; the National and Kapodistrian University of Athens; the Hellenic Institute for the Study of Sepsis; St. James’s Hospital in Dublin; the University of Barcelona; Johns Hopkins University; Emory University; the University of Cincinnati; the University of Florida College of Medicine; Charles University in the Czech Republic; the University of Pennsylvania; the University of Southern California; the Medical College of Wisconsin; the University of Florida College of Medicine–Jacksonville; Washington University in St. Louis; the University of Kentucky; Henry Ford Hospital; Hackensack Meridian Health; Cleveland Clinic; the University of California, Davis; Texas Tech University; the University of South Alabama; Vanderbilt University; the University of Massachusetts; the University of Iowa; the University of Pittsburgh; Geisinger Commonwealth School of Medicine; and Beth Israel Deaconess Medical Center contributed to the research.
Funding for the research came from the Stanford Training Program in Lung Biology, the National Institutes of Health (grants R35GM155165, R21GM150093, R21GM151703, R01HL152083, U19AI167903 and 2U19AI057229-21) and Inflammatix, Inc.
# # #
About Stanford Medicine
Stanford Medicine is an integrated academic health system comprising the Stanford School of Medicine and adult and pediatric health care delivery systems. Together, they harness the full potential of biomedicine through collaborative research, education and clinical care for patients. For more information, please visit med.stanford.edu.
END