
A new era of AI in cancer, pneumonia and blood tests
Recently, a blood test method has been developed that uses artificial intelligence (AI) to detect life-threatening diseases such as ovarian cancer and pneumonia at a very early stage. Daniel Heller’s team, a biomedical engineer at Memorial Sloan Kettering Cancer Center in New York, developed this blood test method using a type of nanotube technology. Nanotubes are tiny tubes of carbon that are about 50,000 times smaller than the diameter of a human hair.
Scientists discovered 20 years ago that nanotubes can emit fluorescent light
More recently, researchers have been able to change the properties of nanotubes to detect the reaction of almost anything in the blood. When nanotubes are placed in the blood, they emit specific wavelengths of light when a substance is stuck to it. But these signals are so subtle that humans fail to understand them. According to Dr. Heller, it is a bit like recognizing a fingerprint, where the pattern of molecules and sensor sensitivity is so subtle that only artificial intelligence can understand it. A machine learning algorithm is used to analyze the nanotube data. This algorithm is trained to determine which samples are from patients with cancer and which are not.
Ovarian cancer is “rare, underfunded and deadly,” says Audra Moran, head of the New York-based charity Ovarian Cancer Research Alliance (OCRA). Like other cancers, the sooner it is detected, the better. Ovarian cancer usually starts in the fallopian tubes, which means it can spread quickly to other parts of the body. Ms Moran said that if ovarian cancer could be detected five years before symptoms appear, mortality rates could be significantly reduced.
Ovarian cancer is a rare disease, so there is a shortage of data to train the algorithm. The researchers initially trained the AI on hundreds of samples from cancer patients. Even then, the method gave good results.
“Our goal is to create a tool that will help doctors detect cancer in gynecological diseases,” she said. She hopes the technology will be ready for general use within three to five years.
In addition to cancer detection, AI has also made significant progress in diagnosing diseases such as pneumonia. California-based company Carius has used AI to identify the exact bacteria causing pneumonia within 24 hours. The company’s CEO, Alec Ford, said that previously, 15 to 20 tests were required to diagnose pneumonia patients, which cost about $20,000. Carius’ technology has made this process much easier and more affordable.
AstraZeneca researcher Dr. Slava Petrovsky has developed an AI platform that can identify 120 diseases
Researchers like Dr. Heller are also using similar pattern-matching technology. However, the lack of access to patient information is still a major obstacle. However, the Ovarian Cancer Research Alliance (OCRA) has taken the initiative to create a large registry of electronic medical records of patients, so that the necessary data can be provided for AI training. OCRA head Audra Moran said, “We are still in the early stages of AI. It is a kind of ‘Wild West’.