Artificial Intelligence is getting better than humans at detecting breast cancer
Since the early 2000s, AI systems have eliminated more than 1.7 million jobs, and that number will only increase as AI improves. Some research estimates that by 2025, AI will eliminate more than 85 million jobs.
But for all the talk about job security, AI is also proving to be a powerful tool in healthcare—specifically, cancer detection. One recently published study has shown that, remarkably, artificial intelligence was able to detect 20 percent more cancers in imaging scans than radiologists alone.
Published in The Lancet Oncology, the study analyzed the scans of 80,000 Swedish women with a moderate hereditary risk of breast cancer who had undergone a mammogram between April 2021 and July 2022. Half of these scans were read by AI and then a radiologist to double-check the findings. The second group of scans was read by two researchers without the help of AI. (Currently, the standard of care across Europe is to have two radiologists analyze a scan before diagnosing a patient with breast cancer.)
The study showed that the AI group detected cancer in 6 out of every 1,000 scans, while the radiologists detected cancer in 5 per 1,000 scans. In other words, AI found 20 percent more cancers than the highly-trained radiologists.
Scientists have been using MRI images (like the ones pictured here) to train artificial intelligence to detect cancers earlier and with more accuracy. Here, MIT's AI system, MIRAI, looks for patterns in a patient's mammograms to detect breast cancer earlier than ever before. news.mit.edu
But even though the AI was better able to pinpoint cancer on an image, it doesn’t mean radiologists will soon be out of a job. Dr. Laura Heacock, a breast radiologist at NYU, said in an interview with CNN that radiologists do much more than simply screening mammograms, and that even well-trained technology can make errors. “These tools work best when paired with highly-trained radiologists who make the final call on your mammogram. Think of it as a tool like a stethoscope for a cardiologist.”
AI is still an emerging technology, but more and more doctors are using them to detect different cancers. For example, researchers at MIT have developed a program called MIRAI, which looks at patterns in patient mammograms across a series of scans and uses an algorithm to model a patient's risk of developing breast cancer over time. The program was "trained" with more than 200,000 breast imaging scans from Massachusetts General Hospital and has been tested on over 100,000 women in different hospitals across the world. According to MIT, MIRAI "has been shown to be more accurate in predicting the risk for developing breast cancer in the short term (over a 3-year period) compared to traditional tools." It has also been able to detect breast cancer up to five years before a patient receives a diagnosis.
The challenges for cancer-detecting AI tools now is not just accuracy. AI tools are also being challenged to perform consistently well across different ages, races, and breast density profiles, particularly given the increased risks that different women face. For example, Black women are 42 percent more likely than white women to die from breast cancer, despite having nearly the same rates of breast cancer as white women. Recently, an FDA-approved AI device for screening breast cancer has come under fire for wrongly detecting cancer in Black patients significantly more often than white patients.
As AI technology improves, radiologists will be able to accurately scan a more diverse set of patients at a larger volume than ever before, potentially saving more lives than ever.
Have you felt a bit like an armchair epidemiologist lately? Maybe you've been poring over coronavirus statistics on your county health department's website or on the pages of your local newspaper.
If the percentage of positive tests steadily stays under 8 percent, that's generally a good sign.
You're likely to find numbers and charts but little guidance about how to interpret them, let alone use them to make day-to-day decisions about pandemic safety precautions.
Enter the gurus. We asked several experts to provide guidance for laypeople about how to navigate the numbers. Here's a look at several common COVID-19 statistics along with tips about how to understand them.
Case Counts: Consider the Context
The number of confirmed COVID-19 cases in American counties is widely available. Local and state health departments should provide them online, or you can easily look them up at The New York Times' coronavirus database. However, you need to be cautious about interpreting them.
"Case counts are the obvious numbers to look at. But they're probably the hardest thing to sort out," said Dr. Jeff Martin, an epidemiologist at the University of California at San Francisco.
That's because case counts by themselves aren't a good window into how the coronavirus is affecting your community since they rely on testing. And testing itself varies widely from day to day and community to community.
"The more testing that's done, the more infections you'll pick up," explained Dr. F. Perry Wilson, a physician at Yale University. The numbers can also be thrown off when tests are limited to certain groups of people.
"If the tests are being mostly given to people with a high probability of having been infected -- for example, they have had symptoms or work in a high-risk setting -- then we expect lots of the tests to be positive. But that doesn't tell us what proportion of the general public is likely to have been infected," said Eleanor Murray, an epidemiologist at Boston University.
These Stats Are More Meaningful
According to Dr. Wilson, it's more useful to keep two other statistics in mind: the number of COVID tests that are being performed in your community and the percentage that turn up positive, showing that people have the disease. (These numbers may or may not be available locally. Check the websites of your community's health department and local news media outlets.)
If the number of people being tested is going up, but the percentage of positive tests is going down, Dr. Wilson said, that's a good sign. But if both numbers are going up – the number of people tested and the percentage of positive results – then "that's a sign that there are more infections burning in the community."
It's especially worrisome if the percentage of positive cases is growing compared to previous days or weeks, he said. According to him, that's a warning of a "high-risk situation."
Dr. George Rutherford, an epidemiologist at University of California at San Francisco, offered this tip: If the percentage of positive tests steadily stays under 8 percent, that's generally a good sign.
There's one more caveat about case counts. It takes an average of a week for someone to be infected with COVID-19, develop symptoms, and get tested, Dr. Rutherford said. It can take an additional several days for those test results to be reported to the county health department. This means that case numbers don't represent infections happening right now, but instead are a picture of the state of the pandemic more than a week ago.
Hospitalizations: Focus on Current Statistics
You should be able to find numbers about how many people in your community are currently hospitalized – or have been hospitalized – with diagnoses of COVID-19. But experts say these numbers aren't especially revealing unless you're able to see the number of new hospitalizations over time and track whether they're rising or falling. This number often isn't publicly available, however.
If new hospitalizations are increasing, "you may want to react by being more careful yourself."
And there's an important caveat: "The problem with hospitalizations is that they do lag," UC San Francisco's Dr. Martin said, since it takes time for someone to become ill enough to need to be hospitalized. "They tell you how much virus was being transmitted in your community 2 or 2.5 weeks ago."
Also, he said, people should be cautious about comparing new hospitalization rates between communities unless they're adjusted to account for the number of more-vulnerable older people.
Still, if new hospitalizations are increasing, he said, "you may want to react by being more careful yourself."
Deaths: They're an Even More Delayed Headline
Cable news networks obsessively track the number of coronavirus deaths nationwide, and death counts for every county in the country are available online. Local health departments and media websites may provide charts tracking the growth in deaths over time in your community.
But while death rates offer insight into the disease's horrific toll, they're not useful as an instant snapshot of the pandemic in your community because severely ill patients are typically sick for weeks. Instead, think of them as a delayed headline.
"These numbers don't tell you what's happening today. They tell you how much virus was being transmitted 3-4 weeks ago," Dr. Martin said.
'Reproduction Value': It May Be Revealing
You're not likely to find an available "reproduction value" for your community, but it is available for your state and may be useful.
A reproduction value, also known as R0 or R-naught, "tells us how many people on average we expect will be infected from a single case if we don't take any measures to intervene and if no one has been infected before," said Boston University's Murray.
As The New York Times explained, "R0 is messier than it might look. It is built on hard science, forensic investigation, complex mathematical models — and often a good deal of guesswork. It can vary radically from place to place and day to day, pushed up or down by local conditions and human behavior."
It may be impossible to find the R0 for your community. However, a website created by data specialists is providing updated estimates of a related number -- effective reproduction number, or Rt – for each state. (The R0 refers to how infectious the disease is in general and if precautions aren't taken. The Rt measures its infectiousness at a specific time – the "t" in Rt.) The site is at rt.live.
"The main thing to look at is whether the number is bigger than 1, meaning the outbreak is currently growing in your area, or smaller than 1, meaning the outbreak is currently decreasing in your area," Murray said. "It's also important to remember that this number depends on the prevention measures your community is taking. If the Rt is estimated to be 0.9 in your area and you are currently under lockdown, then to keep it below 1 you may need to remain under lockdown. Relaxing the lockdown could mean that Rt increases above 1 again."
"Whether they're on the upswing or downswing, no state is safe enough to ignore the precautions about mask wearing and social distancing."
Keep in mind that you can still become infected even if an outbreak in your community appears to be slowing. Low risk doesn't mean no risk.
Putting It All Together: Why the Numbers Matter
So you've reviewed COVID-19 statistics in your community. Now what?
Dr. Wilson suggests using the data to remind yourself that the coronavirus pandemic "is still out there. You need to take it seriously and continue precautions," he said. "Whether they're on the upswing or downswing, no state is safe enough to ignore the precautions about mask wearing and social distancing. 'My state is doing well, no one I know is sick, is it time to have a dinner party?' No."
He also recommends that laypeople avoid tracking COVID-19 statistics every day. "Check in once a week or twice a month to see how things are going," he suggested. "Don't stress too much. Just let it remind you to put that mask on before you get out of your car [and are around others]."
GOOD10: The Pandemic Issue explores big-picture ways that science innovation and communication can usher in a more equitable, more progress-oriented, and safer world.
This issue is a collaboration among GOOD, leapsmag, and the Aspen Institute Science & Society Program.
The GOOD10 format explores fundamental issues facing humanity through the lenses of ten forces pushing the needle toward progress: Places, Philanthropists, Celebrities, Whistleblowers, Companies, Media, Products, Politicians, Scientists, and Actions. Across these categories, we seek to present unexpected and encouraging paradigms emerging from this historic crisis.
This special issue is available as an e-reader version for both desktop and mobile. It is also available as a free downloadable PDF.
TABLE OF CONTENTS:
- PLACES:
55 Lessons Learned About Science Communication Around the World; Quarantining Our Way Into Outer Space - PLACES:
Quarantining Our Way Into Outer Space - PHILANTHROPISTS:
An Exclusive Interview with Wendy Schmidt about Science in the Pandemic Era - CELEBRITIES:
Neil deGrasse Tyson Wants Celebrities to Promote Scientists - WHISTLEBLOWERS:
The Science Sleuths Holding Fraudulent Research Accountable - COMPANIES:
The Biggest Challenge for a COVID-19 Vaccine: Making It Accessible and Affordable - MEDIA:
Isaac Asimov on the History of Infectious Disease—And How Humanity Learned To Fight Back - PRODUCTS:
Will COVID-19 Pave the Way For DIY Precision Medicine? - POLITICIANS:
Will the Pandemic Propel STEM Experts to Political Power? - SCIENTISTS:
Would a Broad-Spectrum Antiviral Drug Stop the Pandemic? - ACTIONS:
Pseudoscience is Rampant: How Not to Fall For It - ACTIONS:
How COVID-19 Could Usher In a New Age of Collective Drug Discovery
THE EVENT:
"The Pandemic Science Summit" focused on how science innovation is key to society's future stability as we emerge from the pandemic, featuring:
Christopher Bailey – Arts and Health Lead, World Health Organization
Elisabeth Bik, Ph.D. – Microbiologist and scientific integrity consultant
Margaret Hamburg, M.D. – Foreign Secretary, National Academy of Medicine; former Commissioner, U.S. Food and Drug Administration
Peggy Oti-Boateng, Ph.D. – Director, Division of Science Policy and Capacity- Building, UNESCO
George Yancopoulos, M.D., Ph.D. – President and Chief Scientific Officer, Regeneron Pharmaceuticals
Kira Peikoff was the editor-in-chief of Leaps.org from 2017 to 2021. As a journalist, her work has appeared in The New York Times, Newsweek, Nautilus, Popular Mechanics, The New York Academy of Sciences, and other outlets. She is also the author of four suspense novels that explore controversial issues arising from scientific innovation: Living Proof, No Time to Die, Die Again Tomorrow, and Mother Knows Best. Peikoff holds a B.A. in Journalism from New York University and an M.S. in Bioethics from Columbia University. She lives in New Jersey with her husband and two young sons. Follow her on Twitter @KiraPeikoff.