The current method doctors use to screen for prostate cancer involves looking at the levels of a protein called prostate-specific antigen (PSA) in a patient’s blood. The problem with a PSA test, however, is that it’s not very accurate. In fact, 70 percent of the people it flags as having prostate cancer don’t.

In search of a better cancer diagnosis solution, scientists at the Korea Institute of Science and Technology (KIST) have developed a form of artificial intelligence (AI) that can make a near 100 percent accurate prostate cancer diagnosis from a urine sample. Rather than search for a single biomarker, the KIST team created an ultra-sensitive sensor that can detect small amounts of four different markers of prostate cancer.

As reported in FreeThink, the researchers collected urine samples from 51 Korean men, with 25 of those men having prostate cancer and the other 26 not having this common form of cancer. The 25 who had cancer contributed two urine samples (one before and one after a rectal exam, which can affect the test results), which means the researchers had a total of 76 urine samples. Then, the researchers used 53 of those urine samples to train two different AIs to look for patterns in the four biomarkers that would indicate that a person has prostate cancer. After this, the researchers tested the AI on the remaining 23 urine samples.

Impressively, one of the AIs correctly categorized every sample it tested, while the other AI produced only one false positive. That’s far more accurate than the conventional PSA test. This is exciting news for the medical world, especially considering that prostate cancer is one of the most common cancers in men, with more than 1.2 million new cases every year. 

Thanks to this new AI from the researchers over at KIST, it could be possible to detect cancer much earlier, thus dramatically improving a patient’s chances of a better health outcome.