Think about how much technology has improved our lives over the years. From the handy smartphone that is capable of tasks previously reserved for a full-sized computer, right in the palm of our hands, to something more grandiose, like self-driving cars. It's all very impressive.
Those innovations are great and life changing, but what about life-saving technology? Don't worry, it's on the way. Google is working on a system that could help save lives.
How is Google helping diagnose cancer?
What we're talking about is a deep learning technology system, also known as Artificial Intelligence (AI), being developed by Google. The system can be used to assist pathologists in detecting cancer in patients.
The AI program is being designed to distinguish between healthy and cancerous tissue. It will also be able to determine if cancer cells have metastasized, or spread, to other areas of the body.
Google researchers said, "Metastasis detection is currently performed by pathologists reviewing large expanses of biological tissues. This process is labor-intensive and error-prone. We present a framework to automatically detect and localize tumors as small as 100 x 100 pixels in gigapixel microscopy images sized 100,000 x 100,000 pixels.
"Our method leverages a convolutional neural network (CNN) architecture and obtains state-of-the-art results on the Camelyon16 dataset in the challenging lesion-level tumor detection task."
What this means is, test results from the AI program will come back quicker and more accurate than traditional tests. The following image is an example of human tissue to be observed on slides.
Image: Example of slides containing human lymph node tissue. (Source: Google)
What were Google's test results?
According to Google, its method yields state-of-the-art sensitivity on the challenging task of detecting small tumors in gigapixel pathology slides, reducing the false negative rate to a quarter of a pathologist and less than half of the previous best result. It further achieves pathologist-level slide-level AUCs in two independent test sets.
Google's method could improve accuracy and consistency of evaluating breast cancer cases, and potentially improve patient outcomes. Future work will focus on improvements utilizing larger datasets.
As of now, this technology is in the development stage, Google hasn't set an implementation time-frame. When it does start being used, it will not replace pathologists because it only searches for cancerous tissue. Pathologists will still be needed to spot any irregularities.
Click here if you want to read Google's full report on its deep learning system.
What do you think about this type of technology, would you welcome it? Leave a comment and tell us your thoughts.