Issue 21: AI, Deep Learning, Social Media, Night Owls, Soda, and VC's
Doctorless AI gets FDA approval
Last week the FDA approved the first artificial intelligence software that can detect diabetic retinopathy without the need of an ophthalmologist. The software, which is called IDx-DR (presumably standing for I diagnose Diabetic Retinopathy), was created by Iowa startup IDx. The FDA wrote in a follow up press release that, “IDx-DR is the first device authorized for marketing that provides a screening decision without the need for a clinician to also interpret the image or results, which makes it usable by health care providers who may not normally be involved in eye care". The device snaps a picture of the patient's retina and sends it to the cloud for algorithmic analysis. The software can detect more than mild diabetic retinopathy 87.4% of the time and had a sensitivity of 89.5%. IDx-DR's approval is not just good for the AI startup scene it's also good for giant tech companies like Microsoft, Google, and IBM. It's estimated that Google has spent $100 million to collect patient data and IBM has spent more than $4 billion to train its medical algorithms.
Deep learning makes cell biology easier and way less boring
If you've ever worked in a biology lab you know that looking at cells underneath a microscope all day can be low-key boring AF. Up until now scientists have had to rely on chemical staining techniques that can damage cell structures. Recently however, doctors at the Gladstone Institute teamed up with the big G aka Google to apply deep learning algorithms to train a robotic microscope to identify cell structures WITHOUT the need for fluoroscopic staining. The deep network can discern whether a cell is alive or dead and get the right answer 98% of the time. This totally blows human researchers out of the water as human can only distinguish alive vs. dead cells with 80% accuracy. The deep learning model can also go one step further and distinguish between a neuronal axon vs. a dendrite which are such similar looking structures that human researchers can't really tell the diff.
Social media playing an unexpected role in medicine
While
much of Congress seems to be convinced that Facebook is purely evil,
doctors and patients alike have found it to be quite useful when it
comes to medicine. Take the example of Allison Ruddick, who turned to
Facebook after being unable to learn more about her stage 3 colorectal
cancer from her doctors. Searching for the hashtags #colorectalcancer
and #nevertooyoung, she stumbled upon a ton of others accounting their
experiences with the cancer, detailing what exactly surgical
interventions involved and even what kinds of ice packs were most
preferable for recovery. Even doctors have benefited from patient
accounts on social media. Dermatologist Bernice Kwong was able to
identify overlooked drug reactions of erlotinib (used to treat lung
cancer) with a few simple google searches. When three of her patients
came to her complaining of hypohidrosis after taking the drug, she
turned to the web and found Inspire.com, where there were almost 5,000
posts from other people claiming to have the same side effect. Seeing
the potential in harnessing patients’ personal accounts of side effects
on social media, Kwong worked with other colleagues to develop an
algorithm for the purpose of processing phrases linking drugs with their
associated adverse reactions. She and her colleagues have just now
published a paper in JAMA detailing the incredible role of social
networks in highlighting uncommon side effects of cancer drugs. So while
Congress may struggle to tell the difference between email, WhatsApp,
and Twitter, physicians at least seem to be getting a better grasp on
social media’s benefits.
Do night-owls die sooner?
In a study published this week in the Journal Chronobiology International, scientists reported that people who sleep later at night had a 10% increase in all-mortality rate compared to “morning people”. Specifically, “night- owls” were found to be more likely to have diabetes, neurological disorders, psychological disorders, gastrointestinal disorders, and respiratory disorders. If you needed an excuse to put off any work/studying for the night and hit the hay, here it is. Catching those extra Z’s now might prevent you from some serious health issues in the future.
Sweet success of cutting out the sugar
In January 2017, Philadelphia instituted a new law that taxed sugary drinks including sodas. The public and soda industry outrage was obviously very predictable, but the result of this controversial tax so far was not as anticipated--it actually worked! Researchers at Drexel University surveyed Philadelphia residents before and after the 1.5 cent per ounce tax on sweetened drinks. In addition, they surveyed people in other comparable cities without the tax. They found that Philadelphia residents were 40% less likely to drink sweetened beverages compared to residents in other cities. What's even better is that, in a study published in the American Journal of Preventive Medicine, water consumption of Philadelphia residents actually increased post-tax. Other cities have recently also adopted a similar strategy including Boulder, Colorado and Oakland, California. The challenge that persists is that it is hard to say whether this decrease in intake will have any measurable health outcomes. For now, this seems to be a promising step in helping the obesity epidemic fizz out.
VC funding for healthcare beats every other industry in 2018
According to the DowJones VentureSource Report, healthcare companies received close to 27% of all VC funding beating every other industry including tech. In total $6.85 billion dollars was raised by these companies which is a 21% increase from the previous quarter and a 4% increase from Q1 2017. Why the increase in funding? Experts cite several reasons including: healthcare costs approaching 1/5 of the US GDP, revolutionary discoveries in CRISPR, and innovations in pharmaceutical discovery. The increased funding comes at an important time in our country's history as we outspend every single country in healthcare costs and yet deliver only mediocre healthcare at best. In fact our primary care outcomes with respect to chronic conditions such as asthma and diabetes are some of the worst in the entire world.
![]() |
IDx's AI software can detect diabetic retinopathy |
Last week the FDA approved the first artificial intelligence software that can detect diabetic retinopathy without the need of an ophthalmologist. The software, which is called IDx-DR (presumably standing for I diagnose Diabetic Retinopathy), was created by Iowa startup IDx. The FDA wrote in a follow up press release that, “IDx-DR is the first device authorized for marketing that provides a screening decision without the need for a clinician to also interpret the image or results, which makes it usable by health care providers who may not normally be involved in eye care". The device snaps a picture of the patient's retina and sends it to the cloud for algorithmic analysis. The software can detect more than mild diabetic retinopathy 87.4% of the time and had a sensitivity of 89.5%. IDx-DR's approval is not just good for the AI startup scene it's also good for giant tech companies like Microsoft, Google, and IBM. It's estimated that Google has spent $100 million to collect patient data and IBM has spent more than $4 billion to train its medical algorithms.
Deep learning makes cell biology easier and way less boring
![]() |
Image from phys.org |
If you've ever worked in a biology lab you know that looking at cells underneath a microscope all day can be low-key boring AF. Up until now scientists have had to rely on chemical staining techniques that can damage cell structures. Recently however, doctors at the Gladstone Institute teamed up with the big G aka Google to apply deep learning algorithms to train a robotic microscope to identify cell structures WITHOUT the need for fluoroscopic staining. The deep network can discern whether a cell is alive or dead and get the right answer 98% of the time. This totally blows human researchers out of the water as human can only distinguish alive vs. dead cells with 80% accuracy. The deep learning model can also go one step further and distinguish between a neuronal axon vs. a dendrite which are such similar looking structures that human researchers can't really tell the diff.
Social media playing an unexpected role in medicine
|
![]() |
Image of hyped owl from the Atlantic |
In a study published this week in the Journal Chronobiology International, scientists reported that people who sleep later at night had a 10% increase in all-mortality rate compared to “morning people”. Specifically, “night- owls” were found to be more likely to have diabetes, neurological disorders, psychological disorders, gastrointestinal disorders, and respiratory disorders. If you needed an excuse to put off any work/studying for the night and hit the hay, here it is. Catching those extra Z’s now might prevent you from some serious health issues in the future.
Sweet success of cutting out the sugar
![]() |
Sodas. |
In January 2017, Philadelphia instituted a new law that taxed sugary drinks including sodas. The public and soda industry outrage was obviously very predictable, but the result of this controversial tax so far was not as anticipated--it actually worked! Researchers at Drexel University surveyed Philadelphia residents before and after the 1.5 cent per ounce tax on sweetened drinks. In addition, they surveyed people in other comparable cities without the tax. They found that Philadelphia residents were 40% less likely to drink sweetened beverages compared to residents in other cities. What's even better is that, in a study published in the American Journal of Preventive Medicine, water consumption of Philadelphia residents actually increased post-tax. Other cities have recently also adopted a similar strategy including Boulder, Colorado and Oakland, California. The challenge that persists is that it is hard to say whether this decrease in intake will have any measurable health outcomes. For now, this seems to be a promising step in helping the obesity epidemic fizz out.
VC funding for healthcare beats every other industry in 2018
![]() |
Image from Dow Jones VentureSource Report |
According to the DowJones VentureSource Report, healthcare companies received close to 27% of all VC funding beating every other industry including tech. In total $6.85 billion dollars was raised by these companies which is a 21% increase from the previous quarter and a 4% increase from Q1 2017. Why the increase in funding? Experts cite several reasons including: healthcare costs approaching 1/5 of the US GDP, revolutionary discoveries in CRISPR, and innovations in pharmaceutical discovery. The increased funding comes at an important time in our country's history as we outspend every single country in healthcare costs and yet deliver only mediocre healthcare at best. In fact our primary care outcomes with respect to chronic conditions such as asthma and diabetes are some of the worst in the entire world.
Our Contributors: