What if you could find prostate cancer not with a biopsy, but with a blood test?
Machine learning powers the data analysis that makes this "liquid biopsy" possible.
Comparing genetic signatures of sick and healthy patients requires analyzing 46,000 data points per patient. At that scale, traditional models are less effective.
Machine learning techniques can help analyze the large genetic datasets powering a prostate cancer blood test.
To make a determination, the test compares genetic expression data to the data of patients known to have cancer.
A "liquid biopsy" is a less invasive way for doctors and patients to learn crucial information.