New Nanophotonic Sample-Testing Chip Could Detect Multiple Viruses or Cancers In Minutes (science.org) 8
Science magazine reports:
Researchers have shown how to conduct thousands of rapid molecular screenings simultaneously, using light to identify target molecules snared on top of an array of tiny silicon blocks. In theory, the tool could be used to spot 160,000 different molecules in a single square centimeter of space. Developed to spot gene fragments from the SARS-CoV-2 virus and other infectious organisms, the technology should also be able to identify protein markers of cancer and small molecules flagging toxic threats in the environment...
"[P]revious sensors have not been able to detect a wide range of target molecules," from very low to very high abundance, says Jennifer Dionne, an applied physicist at Stanford University. In hopes of getting around these problems, Dionne and her colleagues turned to an optical detection approach that relies on metasurfaces, arrays of tiny silicon boxes — each roughly 500 nanometers high, 600 nanometers long, and 160 nanometers wide — that focus near-infrared light on their top surface. This focusing makes it easy for a simple optical microscope to detect the shift in the wavelength of light coming from each silicon block, which varies depending on what molecules sit on top...
[T]he technique could allow doctors to detect viral infections without first having to amplify the genetic material from a patient, Dionne says. Perhaps as important, she notes, an array can be designed to reveal how much target DNA has bound, making it possible to detect in minutes not just whether a particular virus is present, but how intense the infection is. Such information could help doctors tailor their treatments. Current tests can also do this, but they normally take several hours to amplify the genetic material and quantify the results.
Dionne and her colleagues have formed a company called Pumpkinseed Bio to commercialize their new detectors, specifically aimed at detecting minute levels of proteins and other molecules that can't readily be amplified to make them easier to detect. And because only a small number of silicon blocks would be needed to spot individual target molecules, researchers should be able to craft arrays to track a multitude of disease biomarkers simultaneously. "We hope to look at many disease states at the same time," says Jack Hu, a former graduate student in Dionne's lab and head of the new startup. "That's the vision."
Thanks to Slashdot reader sciencehabit for sharing the article.
"[P]revious sensors have not been able to detect a wide range of target molecules," from very low to very high abundance, says Jennifer Dionne, an applied physicist at Stanford University. In hopes of getting around these problems, Dionne and her colleagues turned to an optical detection approach that relies on metasurfaces, arrays of tiny silicon boxes — each roughly 500 nanometers high, 600 nanometers long, and 160 nanometers wide — that focus near-infrared light on their top surface. This focusing makes it easy for a simple optical microscope to detect the shift in the wavelength of light coming from each silicon block, which varies depending on what molecules sit on top...
[T]he technique could allow doctors to detect viral infections without first having to amplify the genetic material from a patient, Dionne says. Perhaps as important, she notes, an array can be designed to reveal how much target DNA has bound, making it possible to detect in minutes not just whether a particular virus is present, but how intense the infection is. Such information could help doctors tailor their treatments. Current tests can also do this, but they normally take several hours to amplify the genetic material and quantify the results.
Dionne and her colleagues have formed a company called Pumpkinseed Bio to commercialize their new detectors, specifically aimed at detecting minute levels of proteins and other molecules that can't readily be amplified to make them easier to detect. And because only a small number of silicon blocks would be needed to spot individual target molecules, researchers should be able to craft arrays to track a multitude of disease biomarkers simultaneously. "We hope to look at many disease states at the same time," says Jack Hu, a former graduate student in Dionne's lab and head of the new startup. "That's the vision."
Thanks to Slashdot reader sciencehabit for sharing the article.
Theranos missed out (Score:3)
If they had talked about nanophotonics and dropped other buzzwords almost no one understands they could have gone a lot longer before getting caught.
Re: (Score:2)
You caught me. I am truly terrified of anything too small for my phone camera's zoom to see.
Unbelievable science claims in medical testing com (Score:2)
Re: (Score:2)
Unbelievable? Well, not by me. I do expect that it will take a lot of development to live up to the advance hype, however. And it may be too expensive to be feasible.
The basic idea is quite reasonable, even though it may not be the best choice for any particular endeavor. I.e. there may (almost) always be better alternatives.
Not Theranos, but not going to work either (Score:5, Informative)
The problem is, this isn't that novel. Technologies like this have been around and published in academia for some time. Hell the Illumina NGS technology is an optical sensor of hundreds of millions of microarrays; that's DNA fragments but applying that to protein chemistry isn't a huge leap. The reason technologies like this don't get traction because they don't have clinical utility. Here's the issue.
No sensor is ubiquitous; there are rather sensitivity windows, or a dynamic range, of what concentration it's optimal to pick up the target. So we can look at 150,000 markers with this; great. Do they ALL have the same level of clinical concentration in the body? No. So let's say a panel of 4-5 markers for one disease is relevant, and another 4-5 is relevant for another disease. The simple fact is, you have those markers in your body right now. Did you know dear Slashdot reader that you have cancerous cells right now in your body? We get them all the time. However it turns out your body is quite good at finding those cancerous cells and eliminating them; it's only when you reach an out of control tipping point that you have cancer. So the problem with sensor sensitivity is just because you pick up a biomarker doesn't mean that a doctor should act on it; if you detect cancer cells do you give someone chemo? Not if the body would clear it on it's own.
So therein lies the crux. Let's say there's two diseases on this test. Most every disease is different and unique: at what level of expression of the marker is the disease at a stage where it's a clinically relevant point to make a decision? What if this test picks up cancerous cells and flu biomarkers. Just because it found the markers doesn't mean you have cancer or the flu, or maybe you do have the flu but not cancer (in a clinical definition, ie a therapy is applied), so now the question is how do you act on the information?
So every biomarker test has to be calibrated to the sensitivity that is relevant to that disease that means that clinical intervention, ie applying a therapy, is now necessary, and every biomarker expression is different. It's not a simple yes/no, but rather at what optimal detection of the biomarker is a positive predictive expression of a clinically defined disease? If the answer is it just detects, then it's not good enough. So the only value in a sensor is does it provide information allowing for a decision to be made, and simple sensitivity just isn't nearly enough to translate to clinical relevance.
So I applaud this group for what they did. It made an interesting paper. But good luck getting it through a diagnostic group like the American Association of Clinical Chemistry; they have seen sensor technology like this hundreds of times and it rarely has any meaning towards applied medicine.
Single cell study? (Score:3)
It would be interesting if they got small enough to implant in a single cell and monitor gene expression at the cellular level. It could give new insights into cellular signaling and wide ranging gene expression processes in a living cell in real time.
Hey, maybe it would need ... (Score:2)
... only one drop of blood!
Is the CEO a cute blond ... I mean, a pathbreaking woman-in-tech CEO?