New Year’s News Smorgasbord
Volume 11, Issue 1 | January 2, 2025
Welcome to 2025, everyone! Once again, we’re starting the year with a buffet of bite-sized bits of news and info. None of us are ready to engage with anything too mentally strenuous this soon after the holiday, don’t you agree? Start with a nibble - or just skip to the end and read about poppy seeds. That’s enough for this week.
Adding AI to screening standard of care could decrease biopsies
The current standard for prostate-cancer screening goes like this:
Research published in European Radiology looked at the second step in this process. The question: Would AI-based analysis of MRI images do a better job than the prostate imaging reporting and data system (PI-RADS) at identifying men who were at low risk of cancer and could therefore avoid having a biopsy?
When used by itself, it didn’t. But when the deep-learning analysis was added to PI-RADS, the researchers saw a big improvement. PI-RADS alone tagged 37% of patients as low risk, with a negative predictive value of 94%. (NPV is the likelihood that someone with a negative test result really doesn’t have the disease.) Adding the deep-learning analysis to PI-RADS bumped the percentage of low-risk patients to 49%, while maintaining the same NPV.
mNGS improves CNS infection diagnosis
Infection in the central nervous system is Bad News, and figuring out which bug is causing the problem is tough. Depending on the technique and type of sample used, the sensitivity of the tests that are currently standard ranges from an underwhelming 46% to a dismal 15%. Fully half the time, clinicians never figure out which microbe caused the illness.
A study in Nature Medicine looked at how well metagenomic next-generation sequencing (mNGS) performed in the clinic over eight years as compared to the traditional tests. This technique works by removing human genetic sequences from the sample, leaving behind only genetic material from the pathogens. While not a complete slam-dunk, it was a big improvement, with an overall “sensitivity, specificity and accuracy of … 63.1%, 99.6% and 92.9%, respectively.”
Lighting up clear-cell renal cell carcinoma
Clear-cell renal cell carcinoma accounts for 90% of deaths due to kidney cancer, but it’s tough to test for. To try to make diagnosis easier, researchers investigated whether they could make the cancer visible on a PET-CT scan.
They injected patients with a radioactive monoclonal antibody that attaches to a protein called CA9. Clear cell kidney cancers have a lot of that protein, but normal tissue has very little. When patients were scanned, areas with a lot of CA9 lit up, indicating that cancer cells were present in those locations. The study, which appeared in The Lancet Oncology, found that the technique was about 86% sensitive and about 87% specific.
Abnormal placental cfDNA result? Screen for cancer.
A small study in the New England Journal of Medicine found that when pregnant people had abnormal results on a test that looks at cell-free placental DNA (cfDNA) in the blood, nearly half of them had previously undetected cancer. (The test is typically used to screen for genetic problems in the fetus.)
For the study, participants with abnormal results underwent an array of cancer-screening tests. Whole-body MRI was the most effective at finding the cancers, which included lymphoma as well as cancers of the colon, breast, lung, pancreas, kidney, and bile duct.
Bird Flu Update
Iowa reports first human case
Multiple deaths in house cats and wild cats
Aged raw cow’s milk cheese to be tested nationwide
The state of Iowa reported its first human case of bird flu on December 20. The patient, who had mild symptoms, was exposed while working with a commercial poultry flock.
Over the last two weeks, several house cats in California and Oregon have gotten infected with H5N1 flu, most as a result of drinking raw milk or eating raw diets. Nearly all of the cats have died.
Twenty of the cats at a wildlife sanctuary in Washington state have died of H5N1 flu. Officials don’t yet know the source of the infection, but they suspect it came from raw meat or from contact with wild birds.
Over the next three months, the FDA is going to test samples of raw cow’s milk cheese from across the US for H5N1 flu. Raw milk cheeses are made from unpasteurized milk. In this country, they must be aged for at least 60 days before being sold, to decrease the chance that they include microbes that can make people sick.
Rats can sniff out more than just TB
Back in June, we described how African giant pouched rats have been trained to detect tuberculosis (TB) in the clinic. These rats are smaller, cheaper, and easier to train and manage than dogs, but with an equally amazing sense of smell. A Frontiers of Conservation Science paper describes how the same nonprofit (APOPO) that trains the rats to detect TB also trains them for explosives detection, landmine clearance, and rescue. (For a humbling and inspiring guide to animals’ sensory abilities, see Ed Jong’s brilliant book, An Immense World.)
Could routine metabolic tests help predict SIDS? Maybe someday.
Routine newborn blood tests gather a lot of information about babies. Could some of that info help tell doctors whether certain babies are at higher risk of sudden infant death syndrome (SIDS)?
A retrospective study published in JAMA Pediatrics this fall answered this question with a qualified “yes.” The researchers’ model evaluated 14 metabolic markers in newborns and combined that data with known SIDS risk characteristics. It was up to 70% successful at identifying infants who would later succumb to SIDS, which sounds great . . . except that it isn’t, really.
SIDS deaths are (fortunately) extremely rare (about 38 cases per 100,000), so the number of false positives the test would generate would overwhelm true positives by several orders of magnitude.
COMMENTARY: Perhaps some useful progress here, but not yet useful in saving lives.
Fun fact: The poppy-seed urban myth is true
When Mara and Liz were kids growing up in New York City, we both heard the same urban legend: If you get screened for heroin after eating a poppy-seed bagel, you’ll test positive. Turns out it’s actually sort of true, if very rare.
Poppy seeds do come from the same plant from which heroin and all other opium-based drugs (opiates) are made. But it’s not the seeds themselves that are the problem. Opiates are made from the milky sap of unripe poppy seed pods, and the seeds can get contaminated with sap when they’re being harvested. So you won’t test positive for opiates after eating just any bagel - it’ll only happen if the seeds on your bagel weren’t properly washed before being sold.
Correction
In our last issue, we reported that the first commercial test for H5 influenza had just been made available. Another commercial test for the disease has been available since October 2024.