Medical scientists in Korea have employed AI-learning to create a new prostate cancer screening that they claim has an accuracy rate of near 100%.
The breakthrough, which is a simple urine strip, could possibly revolutionize testing, as some existing methods can be inaccurate and can result in over-diagnosis and necessitate invasive biopsies.
The current method is a PSA test, which stands for “prostate-specific antigen”. It tests the levels of this particular protein in the blood. This test can have a high misdiagnosis rate.
PSA is produced from both cancerous and non-cancerous prostate cells, and even if the test detects cancerous PSA, there is a risk that it’s diagnosing tumors that would never produce symptoms during a lifetime.
In addition, other conditions such as inflammation of the prostate, an infection, or an enlarged prostate, can also fool a standard PSA test.
Designed at the Korea Institute of Science and Technology, this potential breakthrough utilizes a urine test strip containing an electrical-signal-based ultrasensitive biosensor. AI analysis is also used to quantify the values of four separate prostate cancer factors.
The AI uses an algorithm to determine whether or not the values add up to something that truly indicates cancer. This process led to a very bigh accuracy rate across 76 different tests.