A recent study in The Lancet Oncology showed that AI found 20 percent more cancers on mammogram screens 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.
Miniature lab-grown models of human organs have major advantages over standard animal testing.
But he wasn't the only one to follow this train of thought. In the year since the pandemic began, many researchers have been using organoids to study how the coronavirus infects human cells, and find potential treatments. Beumer's pivot represents a remarkable and fast-emerging paradigm shift in how drugs and diseases will be studied in the coming decades. With future pandemics likely to be more frequent and deadlier, such a shift is necessary to reduce the average clinical development time of 5.9 years for antiviral agents.
Part of that shift means developing models that replicate human biology in the lab. Animal models, which are the current standard in biomedical research, fail to do so—96% of drugs that pass animal testing, for example, fail to make it to market. Injecting potentially toxic drugs into living creatures, before eventually slaughtering them, also raises ethical concerns for some. Organoids, on the other hand, respond to infectious diseases, or potential treatments, in a way that is relevant to humans, in addition to being slaughter-free.
Human intestinal organoids infected with SARS-CoV-2 (white).
Credit: Joep Beumer/Clevers group/Hubrecht Institute
Urgency Sparked Momentum
Though brain organoids were previously used to study the Zika virus during the 2015-16 epidemic, it wasn't until COVID-19 that the field really started to change. "The organoid field has advanced a lot in the last year. The speed at which it happened is crazy," says Shuibing Chen, an associate professor at Weill Cornell Medicine in New York. She adds that many federal and private funding agencies have now seen the benefits of organoids, and are starting to appreciate their potential in the biomedical field.
Last summer, the Organo-Strat (OS) network—a German network that uses human organoid models to study COVID-19's effects—received 3.2 million euros in funding from the German government. "When the pandemic started, we became aware that we didn't have the right models to immediately investigate the effects of the virus," says Andreas Hocke, professor of infectious diseases at the Charité Universitätsmedizin in Berlin, Germany, and coordinator of the OS network. Hocke explained that while the World Health Organization's animal models showed an "overlap of symptoms'' with humans, there was "no clear reflection" of the same disease.
"The network functions as a way of connecting organoid experts with infectious disease experts across Germany," Hocke continues. "Having organoid models on demand means we can understand how a virus infects human cells from the first moment it's isolated." Overall, OS aims to create infrastructure that could be applied to future pandemics. There are 28 sub-projects involved in the network, covering a wide assortment of individual organoids.
Cost, however, remains an obstacle to scaling up, says Chen. She says there is also a limit to what we can learn from organoids, given that they only represent a single organ. "We can add drugs to organoids to see how the cells respond, but these tests don't tell us anything about drug metabolism, for example," she explains.
A Related "Leaps" in Progress
One way to solve this issue is to use an organ-on-a-chip system. These are miniature chips containing a variety of human cells, as well as small channels along which functions like blood or air flow can be recreated. This allows scientists to perform more complex experiments, like studying drug metabolism, while producing results that are relevant to humans.
An organ-on-a-chip system.
Credit: Fraunhofer IGB
Such systems are also able to elicit an immune response. The FDA has even entered into an agreement with Wyss Institute spinoff Emulate to use their lung-on-a-chip system to test COVID-19 vaccines. Representing multiple organs in one system is also possible. Berlin-based TissUse are aiming to make a so-called 'human on a chip' system commercially available. But TissUse senior scientist Ilka Maschmeyer warns that there is a limit to how far the technology can go. "The system will not think or feel, so it wouldn't be possible to test for illnesses affecting these abilities," she says.
Some challenges also remain in the usability of organs-on-a-chip. "Specialized training is required to use them as they are so complex," says Peter Loskill, assistant professor and head of the organ-on-a-chip group at the University of Tübingen, Germany. Hocke agrees with this. "Cell culture scientists would easily understand how to use organoids in a lab, but when using a chip, you need additional biotechnology knowledge," he says.
One major advantage of both technologies is the possibility of personalized medicine: Cells can be taken from a patient and put onto a chip, for example, to test their individual response to a treatment. Loskill also says there are other uses outside of the biomedical field, such as cosmetic and chemical testing.
"Although these technologies offer a lot of possibilities, they need time to develop," Loskill continues. He stresses, however, that it's not just the technology that needs to change. "There's a lot of conservative thinking in biomedical research that says this is how we've always done things. To really study human biology means approaching research questions in a completely new way."
Even so, he thinks that the pandemic marked a shift in people's thinking—no one cared how the results were found, as long as it was done quickly. But Loskill adds that it's important to balance promise, potential, and expectations when it comes to these new models. "Maybe in 15 years' time we will have a limited number of animal models in comparison to now, but the timescale depends on many factors," he says.
Beumer, now a post-doc, was eventually allowed to return to the lab to develop his coronavirus model, and found working on it to be an eye-opening experience. He saw first-hand how his research could have an impact on something that was affecting the entire human race, as well as the pressure that comes with studying potential treatments. Though he doesn't see a future for himself in infectious diseases, he hopes to stick with organoids. "I've now gotten really excited about the prospect of using organoids for drug discovery," he says.
The coronavirus pandemic has slowed society down in many respects, but it has flung biomedical research into the future—from mRNA vaccines to healthcare models based on human biology. It may be difficult to fully eradicate animal models, but over the coming years, organoids and organs-on-a-chip may become the standard for the sake of efficacy -- and ethics.