Getty Images
AI can potentially detect much more from a blood test than is currently possible
This is the third installment of a six-part series examining how AI is changing medical research and treatment.
Ovarian cancer is “rare, underfunded and deadly,” says Audra Moran, director of the Ovarian Cancer Research Alliance (Ocra), a global charity based in New York.
Like all cancers, the earlier it is detected, the better.
Most ovarian cancers start in the fallopian tubes, so by the time they reach the ovaries, they may already have spread elsewhere.
“Five years before presenting a symptom, ovarian cancer may need to be detected to influence mortality,” explains Ms. Moran.
But new blood tests are emerging that use the power of artificial intelligence (AI) to detect signs of cancer in its earliest stages.
And it's not just cancer, AI can also speed up other blood tests to detect life-threatening infections like pneumonia.
Memorial Sloan Kettering Cancer Center
Dr. Daniel Heller trains AI to detect early signs of ovarian cancer
Dr. Daniel Heller is a biomedical engineer at Memorial Sloan Kettering Cancer Center in New York.
His team developed testing technology using nanotubes, tiny carbon tubes about 50,000 times smaller than the diameter of a human hair.
About 20 years ago, scientists began discovering nanotubes that could emit fluorescent light.
Over the past decade, researchers have learned to modify the properties of these nanotubes so that they react to almost anything in the blood.
It is now possible to introduce millions of nanotubes into a blood sample and have them emit different wavelengths of light depending on what sticks to them.
But that remained the question of interpreting the signal, which Dr. Heller likens to finding a match to a fingerprint.
In this case, the fingerprint is a set of molecules binding to sensors, with different sensitivities and binding strengths.
But the patterns are too subtle for a human to distinguish.
“We can look at the data and we won’t find any meaning in it,” he says. “We can only see the different patterns with AI.”
Decoding the data from the nanotubes involved loading the data into a machine learning algorithm and telling the algorithm which samples came from ovarian cancer patients and which came from people who did not. reached.
This included blood from people with other forms of cancer or other gynecological diseases that could be confused with ovarian cancer.
One of the big challenges of using AI to develop blood tests for ovarian cancer research is that it is relatively rare, limiting the data needed to train the algorithms.
And even much of that data is siled within the hospitals that treated them, with minimal data sharing for researchers.
Dr. Heller describes training the algorithm on available data from just a few hundred patients as a “Hail Mary.”
But he says the AI was able to achieve better accuracy than the best cancer biomarkers available today – and that was just the first try.
The system is being further studied to see if it can be improved by using larger sets of sensors and samples from more patients. More data can improve the algorithm, just as algorithms for self-driving cars can improve with more street testing.
Dr. Heller has high hopes for this technology.
“What we would like to do is triage all gynecological illnesses. So when someone comes in with a complaint, can we provide doctors with a tool that quickly tells them that it is more likely that they are “cancer or not, or this cancer rather than that.”
Dr. Heller says it could take “three to five years.”
Warrior
Karius has a microbial DNA database that contains tens of billions of data points.
AI is not only potentially useful for early detection, it also helps speed up other blood tests.
For a cancer patient, contracting pneumonia can be fatal, and because there are approximately 600 different organisms that can cause pneumonia, doctors must perform several tests to identify the infection.
But new types of blood tests make the process simpler and faster.
Karuis, based in California, uses artificial intelligence (AI) to help identify the precise pneumonia pathogen within 24 hours and select the appropriate antibiotic.
“Prior to our test, a patient with pneumonia would have to undergo 15 to 20 different tests to identify their infection during their first week of hospitalization, which represents approximately $20,000 worth of tests,” says Alec Ford, CEO of Karius .
Karius has a microbial DNA database that contains tens of billions of data points. Patient test samples can be compared to this database to identify the exact pathogen.
Mr. Ford says this would have been impossible without AI.
One of the challenges is that researchers do not necessarily understand all the links that an AI could make between the biomarkers tested and the diseases.
Over the past two years, Dr Slavé Petrovski has developed an AI platform called Milton which, using biomarkers from UK Biobank data, can identify 120 diseases with a success rate of over 90%. .
Finding patterns in such a mass of data is something AI can do.
“These are often complex models, where there may not be a single biomarker, but you have to take the whole model into consideration,” says Dr Petrovski, a researcher at pharmaceutical giant AstraZeneca.
Dr. Heller uses a similar pattern matching technique in his work on ovarian cancer.
“We know that the sensor binds and responds to proteins and small molecules in the blood, but we don't know which of these proteins or molecules are specific to cancer,” he says.
More generally, data, or the lack of it, remains a disadvantage.
“People don't share their data, or there's no mechanism to do so,” says Moran.
Ocra funds a large-scale patient registry, with patients' electronic medical records, which allowed researchers to train algorithms on their data.
“It’s still early days – we’re still in the wild west of AI,” Ms Moran says.
More business technology
Source link
