Resistant Fungi are no Fun
Volume 10, Issue 12 | September 26, 2024
In This Issue
Rinse and spit to estimate cancer recurrence
Monitoring brain swelling noninvasively
Dx innovation needed to combat fungal resistance
Bird flu update
New and Noteworthy
Intracranial pressure without drilling
When the brain swells it is a major-league emergency. Unfortunately, because the skull is essentially a sealed box, it’s tough to tell whether the pressure inside it is rising. Right now, monitoring intracranial pressure (ICP) - necessary in cases of head trauma, stroke, or problems with the flow of cerebrospinal fluid - requires drilling a hole in the skull.
But maybe not for much longer. Two independent teams have published AI-based methods that both use invasive arterial blood pressure, photoplethysmography, and electrocardiography - data that’s routinely obtained in these patients - to estimate ICP. While one technique still needs to be validated, the other had a sensitivity of 74% and specificity of 73% on external validation.
To estimate risk of head-and-neck cancer recurrence, rinse and spit
When monitoring head-and-neck cancer patients after treatment, it can be difficult to tell the difference between benign post-treatment changes and a cancer recurrence. A mouthwash-based test could let clinicians know whether they should be concerned.
Saliva collected in the rinse was tested for two biomarkers: total protein level and level of CD44, a tumor-initiating molecule. Post-treatment patients who had either double the normal level of total protein or triple the normal level of CD44 in their saliva were at an estimated 62 - 65% greater risk of recurrence. Research remains in the early stages - the ultimate goal is the development of a rapid test that could be administered in a clinician’s office.
LDTs without Guardrails: Watchdog lawsuit settled
Back in March 2013 EpicGenetics launched a first-of-its-kind, $750 - $1,000 test for fibromyalgia (FM/a), which had been eagerly anticipated by roughly 6% of the US population suffering from the condition. Neurologists were immediately skeptical of the company’s claim of 93% sensitivity and 89% specificity, because the mix of cytokine and chemokine biomarkers were inconsistent with prior published work. Consumer advertising was aggressive: “99% accuracy” was quoted without further explanation. Later, a very similar (identical?) diagnostic was released under the name “100Sure” - it was marketed as a test for so-called “Immune Deficiency Diseases”, an undefined and unrecognized condition (see the STAT review for further details).
These were not the first controversial tests from this company. A 2007 benzene exposure test used to support workers-compensation lawsuits was also criticized, with one academic quoted in STAT, calling it “junk science.”
In 2023 the Center for Science in the Public Interest stepped in and sued EpicGenetics for “false and misleading marketing of laboratory developed tests.” Last month, the company agreed to a settlement, denying wrongdoing but agreeing to some future constraints and voluntarily withdrawing the FM/a and 100Sure tests.
Commentary: Diagnostic tests for complex conditions are urgently needed, with great hopes resting on AI to find truly reliable patterns across multiple biomarkers. However, this particular company’s long track record of playing fast and loose with both science and claims shows the dark side of the current laboratory developed test (LDT) system (see this 2013 Genomeweb article for a historical perspective). It is worth reading the 2023 complaint to see just how far off the rails an LDT test can go.
Bird flu update
Two health care workers who were in contact with an infected individual in Missouri also came down with symptoms. One tested negative; the other recovered before samples could be taken. The infected person had reported no contact with animals, and officials say it may be impossible to determine how that person was exposed.
This week, Nature published a comprehensive review of the state of the H5N1 panzootic (a pandemic for non-humans) that relayed both reassuring and damning messages. Commentary: The reassuring: The odds of H5N1 becoming a human pandemic remain low, because the virus needs a lot of mutations before it can transmit easily from human to human. The damning: We aren’t doing enough to keep track of this thing.
As the report noted, “what keeps scientists up at night is the possibility of unseen chains of transmission silently spreading through farm worker barracks, swine barns, or developing countries, evolving under the radar because testing criteria are narrow, government authorities are feared, or resources are thin. . . . New technologies . . . provide rapid, flexible tools for outbreak response, but are of little use when they are not allowed on the farm.”
Food for Thought
Fungal infection resistance requires diagnostic innovation
We have discussed the issue of pathogen identification and resistance testing before (the Lancet also did a global review of the topic earlier this summer). Plenty of diagnostic approaches enable clinicians to correctly treat patients infected with viruses or bacteria, from the simple and cheap (e.g., using protein C to distinguish virus from bacteria) to the sophisticated and expensive. It’s getting physicians to use them that’s the challenge. (Prescribing a broad-spectrum drug remains quicker and cheaper than ordering and waiting for the results of a diagnostic.)
Correspondence in this week’s Lancet draws attention to the issue of drug resistance in fungal infections. This issue is especially challenging because pathogenic fungal species are all around us and on our skin all the time - they only become a problem when they get out of control.
There are only four classes of antifungal drugs, and they’re quite old. Compared to bacteria and viruses, fungi are genetically closer to humans, so drugs that harm one but not the other are hard to find. But once they’re found, it takes a long time for fungi to develop resistance to them, because fungi cannot share resistance with their “friends and neighbors” through horizontal gene transfer the way bacteria can. Instead, they have to do it the old-fashioned vertical / Darwinian way, one generation at a time. It was only about 20 years ago that fungal multidrug resistance (MDR) was recognized at all.
One fungus that has developed MDR is Candida auris. What is extraordinary is that Candida auris itself was only first discovered in humans recently - in Japan in 2009. In this same period, it has managed to develop sterilization tolerance. The mortality rate for infections with this bug exceeds 35% even in the best hospitals. Indiscriminate crop use of fungicides (e.g., azoles) brought MDR Aspergillus fumigatus into the clinic, as well (see this Nature Review for details). Unfortunately, diagnostic techniques for both fungal identification and fungal susceptibility to drugs are quite limited (as an example, see this CDC article on identification of C. auris).
Commentary: The portfolio of established fungal tests can identify different classes of fungi, but more accurate identification of subspecies and their drug susceptibility still relies on relatively time-consuming techniques (mass spectrometry, NGS, culture). The difficulty of finding effective treatments and the rapid growth of resistance makes better and cheaper diagnostic techniques essential.
Quick Hits
Earlier this week, Mara spoke at the Nephron Healthcare Conference on the Diagnostics Panel highlighting key industry issues including lab consolidation, the growth of multiomics (GSTEP - Genomics, Spatialomics, Transcriptomics, Epigenomics and Proteomics) and the emerging importance of alternative sample types.