One of the key to beating certain diseases and illnesses is by detecting it early. This allows for treatment to begin earlier and also prevents the disease from either spreading or becoming worse. Cancer is one of those instances where early detection can lead to a higher survival rate, and science and technology have helped play a big role in that.
In fact Google seems to be helping with that by working with researchers at the Naval Medical Center San Diego in developing AI that can detect metastatic breast cancer. In fact it seems to work so well that apparently it has a 99% accuracy when it comes to detecting the illness in patients.
This AI is based on Inception-v3, which is an open source image recognition deep learning model. The researchers tested out their AI against the Lymph Node 2016 challenge dataset which contains 399 whole-slide images of lymph node sections from the Radboud University Medical Center and the University Medical Center Utrecht, and like we said, they claim that it managed to score 99.3% in accuracy.
While it wasn’t perfect and would occasionally misidentify certain things, it seemed to be better compared to a practicing pathologist tasked with evaluating the same slides. Now this isn’t the first time that we’re seeing AI used in medicine or cancer. Previously researchers developed AI that can detect skin cancer, prostate cancer, and bowel cancer.
AI won’t be replacing doctors anytime soon, nor should it, but its ability to spot things faster will no doubt help save lives. After all even if it is a false positive, it’s better to be safe than sorry.