According to the National Cancer Institute, over 1.7 million Americans were diagnosed with cancer last year. Aside from the disease itself, cancer patients also suffer from depression, anxiety, and sleep disturbance, which greatly impede their quality of life.
A group of researchers at the University of Surrey have now developed an AI tool to help predict these symptoms. In what is believed to be a first-of-its-kind study, researchers used AI tools to analyze patient data and then predict if the patient would experience depression, anxiety, or sleep disturbance – and to what degree. The AI tool’s predictions were very close to what patients actually experienced.
Researchers are hoping this AI tool can significantly improve cancer patients’ quality of life. If doctors can predict symptoms and their severity before a patient experiences them, they can pre-emptively create a plan to manage, and potentially even prevent, some of them.
This development comes on the heels of another, more significant study, published last month, by a group of researchers at the Osaka University in Japan. These researchers introduced a new AI system to identify different types of cancer cells and determine whether cancer cells are resistant to radiation.
How does it work? To analyze cancer cells, the AI system uses a convolutional neural network (CNN), which you can think of as a series of algorithms that identify complex visual information and classify it correctly without requiring human intervention.
This system is important because a single tumor can have a variety of cancer cells. Having a tool that identifies the cell type and its level of resistance to radiation can help doctors create an effective treatment plan. After all, it’s hard to fight an entity without understanding exactly what one is fighting.
Doctors are not incapable of distinguishing between cancer cells, but the AI system speeds up the process significantly and reduces human error, thus allowing doctors to save more lives.
And the best part? The AI system’s accuracy rate was 96 percent in a 2,000-image test. While not perfect, that number is quite high and means the AI system was able to distinguish between radiation-resistant cancer cells and radiation-sensitive cancer cells the vast majority of the time.
The team at Osaka isn’t done. They are working to establish a universal system to identify and distinguish all cancer types. Here’s to hoping they will be successful!