New diagnostic method applies machine learning to advanced genomic data to detect sepsis

New diagnostic method applies machine learning to advanced genomic data to detect sepsis

Sepsis, the overreaction of the immune system in response to infection, causes an estimated 20% of deaths worldwide and 20% to 50% of hospital deaths in the United States each year. However, despite its prevalence and severity, the disease is difficult to diagnose and effectively treat.

The disease can cause reduced blood flow to vital organs, inflammation throughout the body, and abnormal blood clotting. Therefore, if not recognized and treated quickly, sepsis can lead to shock, organ failure and death. However, it can be difficult to determine which pathogen is causing sepsis or whether an infection is in the bloodstream or elsewhere in the body. And for many patients with symptoms similar to sepsis, it can be difficult to determine whether they even have an infection.

Now, researchers at the Chan Zuckerberg Biohub (CZ Biohub), the Chan Zuckerberg Initiative (CZI), and UC San Francisco (UCSF) have developed a new diagnostic method that applies machine learning to identify advanced genomic data from both microbes and hosts and predict cases of sepsis. As reported in Nature Microbiology on October 20, 2022, the approach is surprisingly accurate and has the potential to far exceed current diagnostic capabilities.

Sepsis is one of the top 10 health problems facing humanity. One of the biggest challenges with sepsis is diagnosis. Existing diagnostic tests are unable to capture the two-sided nature of the disease – the infection itself and the host’s immune response to the infection.”

Chaz Langelier, MD, Ph.D., senior author, associate professor of medicine, UCSF Division of Infectious Diseases and CZ Biohub investigator

Current sepsis diagnostics focus on detecting bacteria by growing them in culture, a process that, according to the researchers behind the new method, is “essential for appropriate antibiotic therapy, which is critical for sepsis survival.” However, culturing these pathogens is time-consuming and does not always correctly identify the bacterium causing the infection. Similarly, for viruses, PCR testing can detect that viruses are infecting a patient, but do not always identify the specific virus that causes sepsis.

“This results in physicians being unable to identify the cause of sepsis in an estimated 30 to 50% of cases,” Langelier said. “This also leads to a mismatch between the antibiotic treatment and the causative pathogen.”

In the absence of a definitive diagnosis, doctors often prescribe a cocktail of antibiotics to stop the infection, but overuse of antibiotics has led to increased antibiotic resistance worldwide. “As physicians, we don’t want to miss a case of infection,” said Carolyn Calfee, MD, MAS, professor of medicine and anesthesiology at UCSF and co-author of the new study. "But if we had a test that could help us determine exactly who doesn't have an infection, then that could help us limit the use of antibiotics in those cases, which would be really good for all of us."

Elimination of ambiguity

Researchers analyzed whole blood and plasma samples from more than 350 critically ill patients admitted to UCSF Medical Center or Zuckerberg San Francisco General Hospital between 2010 and 2018.

But instead of relying on cultures to identify pathogens in these samples, a team led by CZ Biohub scientists Norma Neff, Ph.D., and Angela Pisco, Ph.D., instead used metagenomic next-generation sequencing (mNGS). This method identifies all nucleic acids or genetic data present in a sample and then compares this data to reference genomes to identify the microbial organisms present. This technique allows scientists to identify genetic material from completely different kingdoms of organisms – whether bacteria, viruses or fungi – that are present in the same sample.

However, detecting and identifying the presence of a pathogen alone is not enough for an accurate diagnosis of sepsis, so the Biohub researchers also created a transcriptional profile - which quantifies gene expression - to capture the patient's response to infection.

Genetics & Genomics eBook

Compilation of the top interviews, articles and news from the last year. Download a copy today

Next, they applied machine learning to the mNGS and transcription data to differentiate between sepsis and other critical illnesses to confirm the diagnosis. Katrina Kalantar, Ph.D., senior bioinformatician at CZI and co-first author of the study, created an integrated host-microbe model trained using data from patients diagnosed with either sepsis or non-infectious systemic inflammatory diseases. which enabled sepsis diagnosis with very high accuracy.

“We developed the model by looking at a range of metagenomic data alongside results from traditional clinical testing,” Kalantar explained. First, the researchers identified changes in gene expression between patients with confirmed sepsis and non-infectious systemic inflammatory conditions that appear clinically similar, and then used machine learning to identify the genes that might best predict these changes.

The researchers found that when traditional bacterial culture identified a sepsis-causing pathogen, there was usually an abundance of genetic material from that pathogen in the corresponding plasma sample analyzed by mNGS. With this in mind, Kalantar programmed the model to identify organisms that are present at disproportionately high abundance compared to other microbes in the sample and then compare these to a reference index of known sepsis-causing microbes.

“In addition, we also noted all detected viruses, even if they were at lower levels, because they weren’t supposed to be there,” Kalantar explained. “With this relatively simple set of rules, we were able to do quite well.”

“Almost perfect” performance

The researchers found that the mNGS method and its corresponding model worked better than expected: they were able to identify 99% of confirmed cases of bacterial sepsis, 92% of confirmed cases of viral sepsis, and predict sepsis in 74% of cases with clinical suspicion that had not yet been definitively diagnosed.

“We expected a good performance or even an excellent performance, but this was almost perfect,” said Lucile Neyton, Ph.D., a postdoctoral researcher in the Calfee lab and co-first author of the study. “By using this approach, we get a pretty good idea of ​​what causes the disease and we know with a relatively high degree of certainty whether a patient has sepsis or not.”

The team was also excited to discover that they could use this combined host response and microbial detection method to diagnose sepsis using plasma samples routinely collected from most patients as part of standard clinical care. “The fact that you can actually identify sepsis patients using this widely used, easy-to-collect sample type has big practical implications,” Langelier said.

The idea for the work arose from previous research by Langelier, Kalantar, Calfee, UCSF researchers and CZ Biohub President Joe DeRisi, Ph.D., and their colleagues, in which they used mNGS to effectively diagnose lower respiratory tract infections in critically ill patients. Because the method worked so well, “we wanted to see if the same approach could work in the context of sepsis,” Kalantar said.

Wider implications

The team hopes to build on this successful diagnostic technique by developing a model that can also predict the antibiotic resistance of pathogens discovered using this method. “We've had some success with this in respiratory infections, but no one has found a good approach for sepsis,” Langelier said.

In addition, researchers hope they will eventually be able to predict outcomes of patients with sepsis, "such as mortality or length of hospital stay, which would provide important information that would allow physicians to better care for their patients and target resources to the patients who need them most," Langelier said.

“There is great potential for novel sequencing approaches like this to help us more accurately identify the causes of a patient's critical illness,” Calfee added. “If we can do this, it will be the first step towards precision medicine and understanding what is going on at the individual patient level.”

Source:

Chan Zuckerberg Biohub

Reference:

Kalantar, KL, et al. (2022) Integrated host microbial plasma metagenomics for sepsis diagnosis in a prospective cohort of critically ill adults. Natural microbiology. doi.org/10.1038/s41564-022-01237-2.

.