UCSF finds some success with DNA test for infections
A tool designed at UCSF that uses DNA sequencing to identify hard-to-detect infections in the brain and spinal cord proved more useful than traditional diagnostic tests in a large study, but it’s still too expensive and cumbersome to make it practical for widespread use, infectious disease experts said.
The study of 204 patients was published Wednesday in the New England Journal of Medicine. All of the patients had been hospitalized, most at UCSF, in serious condition with inflammation of the brain or spinal cord. In most cases, the source of illness was never found.
Among the 58 patients diagnosed with a viral, bacterial or other infection, the new sequencing tool proved the most reliable of all diagnostic tests, according to the study. And the tool was able to identify 13 infection-causing pathogens that would have gone undetected by traditional testing.
In eight of those 13 cases, finding the cause of infection changed the course of treatment — in one case, a woman likely was spared having to get a liver transplant when the test found that the source of her illness was a rare but treatable viral
infection.
“If this test were not available, we would have missed more than onefifth of the infections that were diagnosed,” said Dr. Charles Chiu, director of the UCSF-Abbott Viral Diagnostics and Discovery Center and senior author of the study. “These are critically ill individuals. For eight of them, we were able to modify their therapy and treat them appropriately.”
Meningitis, encephalitis and myelitis are all potentially deadly conditions in which the brain or spinal cord become inflamed due to an attack on the body by its own immune system or infection by an outside pathogen like a virus. In the UCSF study, nearly half of the patients were treated in the intensive-care unit, and many were people with weakened immune systems that leave them especially vulnerable to infection.
Determining the cause of inflammation is notoriously challenging, in part because the symptoms of different conditions often overlap, and finding a source of infection requires analysis of samples taken from spinal cord fluid or the brain.
Traditionally, doctors diagnose infections in these cases by putting together a list of the most likely causes — particular viruses or bacteria, for example — and then sending the fluid or tissue sample to a lab to test for those pathogens. But that process may require multiple rounds of testing that still come up with nothing. If the cause of infection is a rare pathogen, then it will probably go undetected because doctors won’t think to look for it.
The UCSF tool takes the same sample and sequences all of the DNA in it. After removing known human genetic material, scientists run the sample through a database with DNA sequences for more than 10,000 pathogens. In the study, the tool was tested on people with brain and spinal cord inflammation because those patients tend to be the sickest and most difficult to diagnose. But the tool can be used on blood and other tissue samples.
“The traditional paradigm is to think of what are the most likely possibilities, then do test after test after test until you find something,” Chiu said. “Now we can do everything at once.”
Scientists in the study were able to diagnose one patient with St. Louis encephalitis virus, a mosquito-borne pathogen that hadn’t been seen in California since 1986 and that doctors never would have considered. The patient, a man from Bakersfield, eventually died from that virus, but diagnosing him correctly helped refine his care and informed public health officials of a potential resurfacing threat, Chiu said.
Doctors also used the tool on a woman hospitalized with brain and spinal cord inflammation who was experiencing liver failure. They had not been able to determine the cause of illness and were preparing her for a liver transplant when the sequencing tool came up with hepatitis E, which is curable. The woman did not need the transplant.
Several infectious disease experts said they’ve been interested in the potential of the UCSF sequencing tool and said that it’s clearly proved useful in specific cases. Whether it will be effective on a wider scale remains to be seen, though. The tool is much more expensive than traditional tests — about $2,200 compared to $50 to $500 — and it takes several days to produce results.
And though it appears to be more sensitive than traditional testing, the tool still didn’t help scientists make a diagnosis in the bulk of cases in the new study.
“In my mind, there had been hope that this methodology would be better than it seems to be. But it clearly provides diagnoses when other means fail,” said Dr. Stan Deresinski, an infectious disease specialist at Stanford Health Care. “It would be more applied to circumstances where at least your initial round of testing doesn’t give you an answer.”
Chiu acknowledged the limitations of the tool and said his team is working on strategies to improve speed and affordability.
“My goal is to to develop this test to be routine,” he said. “Right now, this will primarily be applied in the sickest patients. But this is a transformative technology.”